Browserless tools and the browser tool share one browser per session, and now run one at a time instead of racing on the same page. browserless_content, browserless_scrape and browserless_screenshot read in a tab of their own, so they no longer move the page the agent is working in.

A browser key press now reports where focus landed (Pressed Tab. Focus is on link "Vision".), and read_page marks the focused element, which makes keyboard and accessibility testing practical.

Connecting a connector that is turned off for the organization now says so, and where an admin turns it on, instead of reporting an unknown connector.

In code: routes under /v1/channels/{channel_id} start sessions, send messages, follow events over SSE, answer questions, approve tools and cancel turns. A key sees only the sessions it started. Channel settings add end users through your own identity provider, browser origins, Idempotency-Key retries, a daily spend limit per caller, and access for organization members.

See Call your agent from code, Agent API and #4438, #4490, #4508, #4511, #4514, #4516.

tools github list-pulls '{"repo":"acme/api","state":"open"}' | jq -r '.[].number'

Every call goes through the turn’s tool policy. A script that would make a call needing approval is held before it starts. tools plan lists a script’s calls and their risk without running anything. Tools with no risk hints are rated once, which Rate unlabeled tools can turn off. Each call shows in the session timeline under its shell call.

See Tools in Shell and #4385, #4388, #4435, #4439, #4443, #4501.

In the UI: a session now opens on Trace, which replaces the Transcript and Timeline tabs. A turn rail with a minimap leads to each turn’s steps: model calls, tool calls, approvals, sub-agents and messages, with waterfall timing bars. The step inspector shows input, output, the full model request and raw events. Batches and sub-agents expand inline, search jumps to a hit, and j and k move between steps. The view, error filter and selected step live in the URL.

In code: read the same data from /v1/sessions/{id}/trace, or with everruns sessions trace get and everruns sessions trace turns list.

See Web UI, the sessions CLI reference and #4481, #4500, #4503, #4556.

The server now keeps an in-memory SQLite cache per database and, after every write, stores only the changed 4 KB pages in PostgreSQL in one transaction. A server that does not have the database cached rebuilds it from the stored pages, and each commit is a compare-and-set so concurrent writers cannot overwrite each other. Agents keep using sql_execute, sql_query and sql_schema as before.

See SQL database and #4553.

{ "ref": "poppy_client", "config": {} }

The deployment holds one personal agent identity with pairwise user IDs, and keys and session tokens stay in the control plane. A Poppy channel type serves the company side of the protocol.

See Poppy personal agent and #4479, #4549, #4558.

For agents: with the User MCP Servers capability (user_mcp), each turn adds the enabled servers of the person chatting. Platform Chat has it on, with manage so it can add and connect servers mid-conversation after approval. Personal servers load their tools on demand unless you turn off Load tools on demand.

See User MCP servers, the capability and #4295, #4315, #4345, #4381, #4496.

In the UI: open Settings > MCP catalog and click Tools on a server. Suggest labels asks a decision model to judge unlabeled tools; nothing changes until someone picks a choice or clicks Use suggestion.

In code: PUT /v1/mcp-servers/{id}/tools/{tool_name}/label with {"label": "read_only"}, or everruns mcp-servers label-tool and suggest-labels on the CLI. Label changes are in the server’s history.

See Which tools ask before they run and #4475, #4498.

Organizations are migrated once: agents bound to the old Worker move to Bashkit Worker or Sandbox Worker based on their sandbox policy. Worker Base and Generic stay as deprecated harnesses for existing bindings.

See Built-in harnesses and #4484.

For agents: llm_judge and mcp checks are now valid on the output stage. They run once on the finalized assistant message, with the same caps, advisory downgrade and fail-open behavior as the tool stages.

See Guardrails and #4459, #4542.

Turn on Only when the CI result changes in the GitHub trigger form to wake the agent only when the outcome for a workflow and branch changes, for example from passing to failing, instead of on every run. Triggers expose this as the changes_only filter.

See Agent triggers, Summarize GitHub pull requests and #4541, #4546.

This fits Slack channels and conversations with several people, where the agent should answer only when it has something to add. The setting is on the agent in the web UI, the API, agent packages, the Framework builder and the CLI, and it replaces Slack’s per-endpoint reply mode. Every sent message is recorded as a conversation.message event.

See Explicit communication and #4447.

Every outbound request is logged on the everruns::egress::audit target with its decision, host, status, organization and session.

A platform administrator can let an organization extend the allowlist with up to 50 host patterns, or grant it open egress. Manage this with everruns orgs egress-allowlist.

See Network access, the orgs CLI reference and #4432, #4437, #4452, #4534.

Connecting a personal ChatGPT plan used to depend on an organization feature flag that hosted deployments kept off. The flag is gone: connecting the provider is the opt-in, and each connection already belongs only to the user who signs in.

Open Models → Providers → Connect provider → ChatGPT → Sign in with ChatGPT, approve plan use in ChatGPT, then pick a discovered model in the chat model picker. The composer shows Using your ChatGPT plan. Other users, shared defaults and Playground cannot use your connection.

See ChatGPT plan and #4329, #4547.

In code: POST /v1/sessions/{id}/messages accepts an optional client_message_id. Sending the same one again returns the stored message instead of starting a second turn, and the response reports whether the message started, steered or resumed a turn. The CLI takes --client-message-id.

See Chat and threads and #4502, #4494, #4480.

Clients must move to the Agent API routes under /v1/channels/{channel_id}. Old request shapes get 400 api_endpoint_retired pointing at the Agent API docs, and /v1/apps/{app}/api/{chan} routes return 410.

See Call your agent from code, Agent API and #4527.

This is the Require reasons from agents feature flag, now on by default. An organization can turn it off to let such changes through with a warning, which the everruns_entity_changes_without_reason_total metric counts. People in the UI and scripts calling the API may still leave the reason out.

See Reasons and #4271, #4552.

Budget summaries also carry the budget ID, so a client can tell which of several stacked budgets stopped the turn. To continue, top up the exhausted budget with the budget API or CLI.

See Budgets and #4477.

In the UI: an organization administrator turns on Voice in Settings → Features. Then add a voice channel from the agent’s Integrations tab and use its talk button. Speech uses an OpenAI provider.

In code: everruns::voice puts a voice channel in front of any Framework session, and serve’s voice feature gives every top-level agent browser calls.

See Voice, Framework voice and #4394, #4402, #4406.

Providers moved out of Settings into Registries → Models, which has three tabs. Models lists every model from every connected provider and flags new ones a sync discovered for review. Providers connects a provider in one sheet: driver, name and key, then the models to enable, with recommended ones ticked. Defaults sets the organization’s default model, decision model and default providers per service.

Provider names are now unique within an organization. Managing providers and models needs the org:models:manage permission, which replaces the separate provider permission.

See Models, Providers and #4458.

• Codex session reliability: Legacy Codex sessions keep host observability metadata and the session HTTP header while omitting unsupported request-body metadata, fixing ChatGPT sessions that failed with HTTP 400 (#4464).

• Agents home and a New agent page with an agent builder
• Fix ChatGPT plan sessions failing with HTTP 400
• See and revoke AI clients connected to Everruns
• Models and providers move into one registry page
• Voice channels let people call an agent

See the GitHub release and the full release notes for every pull request in this version.

• Agents home: Enable the Agents home feature for run activity, setup alerts and a shared Channels view (#4399).
• Agent scripts: Save and manage reusable shell scripts on agents through the API, CLI and MCP, then run them with tools scripts using the caller’s tools and identity (#4390, #4400).

• Save shell scripts on agents and run them from triggers

See the GitHub release and the full release notes for every pull request in this version.

• Tools in Shell: Call visible tools from a Bashkit script, load deferred MCP servers, and pause for approvals before the script starts (#4385, #4386, #4388).
• Browser tools: Provider-neutral Browserless browser tools and a native Claude browser toolset adapter (#4375, #4376).
• Coordinator threads: Coordinate child threads in Framework applications and Platform Chat (#4377, #4372).

• Browser tool for agents on Browserless
• Coordinator threads in Platform Chat and the Framework

See the GitHub release and the full release notes for every pull request in this version.

In the UI: open Settings → My agent experience to see each client and when it was last used, and revoke any you no longer use. Revoking deletes the client’s refresh tokens, and its next request is refused.

In code: GET /v1/user/connected-clients lists them and DELETE revokes one. Every MCP request checks that its grant is still live.

See Use Everruns in AI tools and #4461.

Everruns now leaves that metadata out of the request body while keeping the session HTTP header and host observability metadata. No configuration change is needed; existing ChatGPT connections work after upgrading.

See ChatGPT plan and #4464.

In the UI: the Agents page shows each agent’s channels and triggers, 24-hour run bars and a Needs attention list for setup problems. Its Channels view lists every channel in the organization with publish and unpublish, people reached and fastest first reply. It replaces the Exposures page; /exposures now redirects there.

New agent offers Describe, Example, Blank and Import. Describe drafts the agent with the utility model, and Create makes the agent with its draft channels and schedule trigger.

In code: everruns agents draft runs the same agent builder from the CLI.

See Web UI, Channels and #4399, #4448, #4457.

Inside the shell, tools scripts lists and runs them with the caller’s tools and identity, and with manage_scripts on the agent can save new ones.

A schedule or webhook trigger can run a script without asking the model:

{ "script": { "script": "triage-prs", "input": { "pr": "{{payload.pull_request}}" } } }

Input strings are templates over the event, so a webhook payload reaches the script.

See Saved scripts, Agent triggers and #4390, #4400, #4403, #4423.

• A2A 1.0 HTTP+JSON and PACT: A2A channels serve the HTTP+JSON binding, the PACT Identity profile, and PACT Delegated sign-in, step-up and receipts (#4322, #4328, #4350, #4352)
• AgentID sign-in: Public chat and channels accept AgentID, and agents sign in with their AgentMail inbox (#4334, #4326, #4337)
• MCP in conversations: Users manage their own MCP servers, connect them mid-conversation, and deferred servers load tools on demand (#4315, #4343, #4345, #4349)
• Claude Haiku 5.5 and Daytona desktop computer use (#4359, #4365)

• Agent versions are retired in favor of change history
• A2A channels serve HTTP+JSON and the PACT profiles
• Sign in to channels and Public Chat with AgentID
• Claude Haiku 5.5 model added for Anthropic providers

Before upgrading, read the breaking changes: Agent versions are retired in favor of change history.

See the GitHub release and the full release notes for every pull request in this version.

Agents can now operate a full Linux desktop instead of only a browser page. The experimental computer_use_desktop capability runs the provider-neutral computer tool on an E2B desktop sandbox, and computer_use_daytona runs it on a Daytona sandbox through Daytona’s Computer Use API. The session creates the desktop on first use, keeps it between calls and deletes it when its lease ends.

The computer tool also gained zoom, cursor position, mouse button down and up, hold-key and modified drag actions on Daytona, E2B and Browserless.

Connect E2B or Daytona first, then add one computer-use capability to the agent; only one can provide the computer tool.

See Computer use and #4244, #4365, #4367.

In code: the Framework’s everruns::coordination capability runs threads as sessions of the same engine. Threads wake the coordinator with automatic updates, and coordination::threads returns each thread’s status, such as NeedsYou, Working or ReadyForReview.

See Chat and threads, Coordination and #4372, #4377.

The browser_use capability gives an agent a browser tool on a Browserless browser. Pages are read as an accessibility tree with element refs, and the agent acts by ref or viewport coordinate, fills forms, finds text and works across tabs. Traffic goes through the Everruns network guard.

On Claude models that support Anthropic’s native browser toolset, the capability uses it directly; set native_tools: false to keep the function tool.

Browser use is experimental and ships behind the browserless_browser_use feature flag.

See Browser use and #4375, #4376.

See the GitHub release and the full release notes for every pull request in this version.

Its profile has tiered pricing, up to 128K output tokens, always-on adaptive thinking with medium effort by default, preserved thinking, tool search and the computer toolset. Select it in an agent’s model picker or reference claude-haiku-5-5 wherever a model ID is accepted.

See Anthropic and #4359.

To let agents sign in to a Public Chat channel, register an AgentID client, set AGENTID_CLIENT_ID and AGENTID_CLIENT_SECRET on the server, and set the channel’s sign-in to AgentID. The chat page then shows Continue with AgentID.

Going the other way, the experimental AgentID sign-in capability gives an agent the agentid_authorize tool, which approves an app’s AgentID sign-in for the agent’s AgentMail inbox.

See Sign in with AgentID, AgentID and #4326, #4334, #4337.

Add pact to an A2A channel’s config to accept personal agents under the PACT 1.0 Identity profile, with each conversation bound to the calling user. With delegation on, the channel runs an OAuth device-code sign-in so a personal agent can act as a company user through the configured MCP server, and delegated replies carry a signed receipt.

See A2A, Personal agents (PACT) and #4322, #4328, #4350, #4352.

• Agents get avatars from presets or an uploaded image
• Sandbox Templates and a Sandboxes fleet view
• Change history with reasons and restore for every entity

See the GitHub release and the full release notes for every pull request in this version.

In the UI: open the ⋯ menu on an entity’s page. History lists changes with their reasons, and Restore this point asks for a reason. Manager notes holds context agents read before changing the entity.

In code: pass a reason with --reason on the CLI or the Everruns-Change-Reason header, and read history with everruns history list, show, diff and restore.

See Change history, the history CLI reference and #4228, #4264, #4272.

Version endpoints, CLI verbs and fields such as agent_version_id and default_version_id are removed. To go back to an earlier configuration, use History → Restore this point on the agent page or everruns history restore.

See Agent revisions replace agent versions and #4312.

In the UI: open Sandbox Templates and select New Sandbox Template. The Sandboxes page lists every sandbox across providers and states, with a lifecycle timeline, a Needs attention filter and pause, resume and delete in the detail drawer. Both sit in the collapsible Sandboxes sidebar group.

In code: GET /v1/sandboxes lists the fleet; /stats and /timeline back the page.

See Sandbox Templates and #4199, #4246, #4262.

In code: PUT /v1/agents/{agent_id}/avatar uploads an image as the multipart field file, and PUT /v1/agents/{agent_id}/avatar/preset selects a preset. The agent’s avatar field lists public, immutable image URLs. The A2A Agent Card carries it as iconUrl, and one-click Slack apps use it as their icon.

See Web UI and #4198, #4220.

• Framework crates consolidated into everruns-core and the facade

Before upgrading, read the breaking changes: Framework crates consolidated into everruns-core and the facade.

See the GitHub release and the full release notes for every pull request in this version.

Everruns 0.40.0 is a maintenance release with fixes and internal changes and no user-facing entry of its own.

See the GitHub release and the full release notes for every pull request in this version.

• Chat threads keep conversations and ongoing work beside Chat
• Slack permission issues appear in Health with recovery steps

See the GitHub release and the full release notes for every pull request in this version.

Review issues in Settings > Health, beside the channel’s settings, and under Action required in the notification menu. Managed apps get a repair that extends the app’s manifest before Slack asks for consent again; manually configured apps get scope, reinstall and token guidance. Reading or snoozing an issue keeps it active until a fresh check passes. From the CLI, use everruns health-issues.

See Resolve installation issues. PR #4134.

Chat now always uses the managed Platform Chat agent. The sidebar opens your permanent conversation, New chat starts a side conversation, and All chats lists only your personal conversations. Platform Chat runs on the shell-based implementation for every organization.

See Chat and Threads and Platform Chat. PRs #4060, #4076, #4121, #4129.

Upgrade your configuration:

• FEATURE_*=preview becomes internal.
• true becomes adoption to keep organization opt-in, and false becomes off.

Invalid values, including the old ones, disable the feature. The separate DEPLOYMENT_GRADE=preview is unchanged.

See Environment variables. PRs #4115, #4137.

• Personal ChatGPT plan connections: Self-hosted deployments can enable personal account sign-in and discover its available models; connections stay scoped to the signed-in user (#4136).
• Stock subagents for custom hosts: Library hosts can run the stock subagent and handoff tools through a store of portable runtime definitions, without constructing control-plane records (#4078).

• Connect a personal ChatGPT plan as a model provider
• Conversation becomes the default harness and Generic is deprecated
• Feature flag grade preview renamed to internal for operators
• Playground for testing agents, and Chat as one permanent conversation

Before upgrading, read the breaking changes: Conversation becomes the default harness and Generic is deprecated, Feature flag grade preview renamed to internal for operators.

See the GitHub release and the full release notes for every pull request in this version.

• Approval history in context: Chats show approval requests and recorded consent; organization admins and owners can filter existing spoken-consent audit history (#4110).
• Agent channels at a glance: Compact agent cards show transport and live, draft or disabled status, with links to Integrations (#4109).
• Navigation keeps your place: Sidebar selection updates as navigation starts, and agent tabs survive refresh through their URL (#4108, #4111).
• Slack installation returns to setup: Valid OAuth callbacks return to the owning agent’s endpoint editor; invalid callbacks use the agents page (#4107).

• Approval history in chat and a new Approvals audit page

See the GitHub release and the full release notes for every pull request in this version.

• Choose an Agent Environment: Author named Bashkit or Daytona profiles and select the profile when starting a Chat or Playground session.
• Slack installation repair: Add to Slack works for draft endpoints, requests declared bot scopes, and shows compact errors with retry.
• Session storage permissions: Listing stored values or secret names requires session view; denied callers receive no private data.

See the GitHub release and the full release notes for every pull request in this version.

• Safer sandbox tools: HTTP egress fails closed on empty DNS pins, and forged session secrets cannot redirect Docker tools (#4058, #4061).
• Interactive questions in chat: AskUser runs through interactive turns so agents can ask for input while working (#4044).
• Platform Chat for every organization: the shell-powered Platform Chat harness becomes the default for every organization (#4060).

• Computer use capability lets agents operate a browser screen

See the GitHub release and the full release notes for every pull request in this version.

Generic is deprecated but unchanged. Selectors hide it behind Show deprecated. On upgrade, agents that inherited Generic as the default are pinned to it, so their behavior does not change. A custom harness whose name collides with a new built-in keeps its ID and definition, but its slug moves to <name>-custom; update callers that look it up by name.

In code: Harness::base(), conversation(), worker_base() and worker() build the same presets.

See Harnesses and the upgrade notes. PR #4133.

On models with a native computer tool, the agent uses the provider’s tool instead; set native_tools to false to always use the function tool. max_actions_per_session caps actions per session, and navigation passes the egress guard. No call waits for a person by default: add soft_approval to record confirmations, or tool_approval to approve every call.

See Computer use. PRs #3922, #3968, #4064.

ChatGPT now appears in the provider grid. Open Models > Providers > Connect provider > ChatGPT > Sign in with ChatGPT, approve plan use in ChatGPT, and return to Everruns. The connection belongs to you, not the organization: its models are available only to you, and the detail page shows account status, quota guidance, and reconnect and disconnect controls.

Self-hosted deployments use OpenAI’s Sign in with ChatGPT for open-source applications. The ChatGPT plan transport is now part of everruns-drivers.

See ChatGPT plan. PR #4136.

Create a New thread, search conversations, and open work from its card, where you can answer pending input or stop it. Resolve thread puts a conversation away without losing history, and Reopen thread continues it. The panel becomes a drawer on mobile. Turning the feature off restores the previous Chat and keeps conversations and their URLs.

See Chat and Threads. PR #4146.

Organization admins and owners also get an Approvals page under Operational. It filters all activity, recorded approvals, or requests, and links each entry back to its session. It reads the existing approval audit records, so earlier history appears too.

See Soft approval. PR #4110.

• Execution kernel in a JavaScript isolate: the engine’s execution kernel runs inside a JavaScript isolate, with a celld engine cell (#4024)
• Durable serve apps on celld: new everruns-serve-celld hosting target snapshots and restores a serve app across process and node loss (#4019)
• Workers AI model discovery: the Cloudflare driver lists Workers AI models, and Chat Completions streams keep tokens sent as JSON numbers (#4021)

• Cloudflare AI Gateway and Vercel AI Gateway model providers
• Run an Everruns serve app durably on celld

See the GitHub release and the full release notes for every pull request in this version.

• everruns-durable workflow engine published on crates.io
• Host Everruns serve apps on Amazon Bedrock AgentCore Runtime

See the GitHub release and the full release notes for every pull request in this version.

• AG-UI 1.0: A new everruns-ag-ui crate carries the AG-UI 1.0 wire types (#3964), and the AG-UI endpoint moves to protocol 1.0 with interrupts, resume, token usage, frontend tools, subagents and run metadata (#3965, #3967, #3972, #3979). Any framework session serves AG-UI behind the ag-ui feature, with a durable thread store and a ready-made axum handler (#3976, #3998, #4000), everruns-serve exposes an AG-UI channel (#3978), and a consumer pipeline and SSE client ship with an upstream conformance corpus (#3982).
• A2A 1.0: The A2A endpoint speaks A2A 1.0 and returns task results (#3999), with ListTasks, SubscribeToTask (#4010) and push notifications (#4013); everruns-serve serves every agent over A2A behind an a2a feature (#4014).
• Amazon Bedrock AgentCore hosting: The new everruns-serve-agentcore crate hosts agents on AgentCore Runtime, with Bedrock through the execution role, WebSocket transport and durable approvals (#3990, #4012).
• OpenAI Agents API backend: An opt-in durable OpenAI Agents API backend (#3957) projects events, usage and cost (#3961), supports session lifecycle and portability (#3963), and enforces Everruns policy at tool boundaries (#3959).
• Native computer use with hard approval: Native OpenAI and Anthropic computer tools run behind a hard tool-approval gate (#3968, #3956).
• Slack workspaces per organization: Organizations connect Slack workspaces and pick one per agent (#3977).
• Agent version pinning: Endpoints and triggers can pin a specific agent version (#3975).
• More ways to reach models: Cloudflare and Vercel AI Gateway drivers (#3996), an opt-in Responses WebSocket transport for OpenAI (#3981), and OpenAI background responses that re-attach after a worker restart (#3969).
• GitHub and MCP triggers: PR reviewer and security scanner templates on the GitHub App (#3980), and agents that wake on inbound MCP events (#3970).
• everruns-durable on crates.io: The durable workflow engine fires timers, runs child workflows and ships with a bundled PostgreSQL schema (#3994, #4008).

• PR reviewer and security scanner agent templates for GitHub
• MCP Events add session webhooks and triggers on MCP server events
• OpenAI structured output and background mode for long reasoning
• Connect several Slack workspaces per organization
• A2A 1.0 endpoint with task results, streaming and push notifications
• AG-UI 1.0 endpoint and AG-UI streaming for Framework sessions

See the GitHub release and the full release notes for every pull request in this version.

Two gateway providers are available. Cloudflare reaches many upstream providers and Cloudflare’s own Workers AI models through one account token and account ID. Vercel AI Gateway uses its OpenAI-compatible Responses endpoint and syncs its model list.

The Cloudflare provider lists Workers AI text-generation models with their context window and tool support. Third-party model IDs on Cloudflare are still added by hand, because no endpoint lists what an account can bill.

This also fixes a token dropped from Cloudflare streams when the model emitted a purely numeric token.

See Cloudflare and Vercel. PRs #3996, #4021.

In code: the ag-ui feature lets any Rust host stream a Framework session to an AG-UI client without the Everruns server. Session::ag_ui(input) returns the run as AG-UI 1.0 events, and ask_user questions and tool approvals become interrupts. A durable thread store keeps history between runs, and the ag-ui-axum feature adds a ready-made route:

cargo add everruns --features ag-ui-axum

See AG-UI channels and Serve AG-UI. PRs #3965, #3967, #3972, #3979, #3976, #3998, #4000.

In code: the a2a feature of everruns-serve answers A2A 1.0 for every top-level agent at POST /v1/e/{agent}/a2a, the same paths an Everruns A2A endpoint uses.

See A2A and Framework A2A. PRs #3999, #4010, #4013, #4014.

Each agent you put in Slack gets its own Slack app. In the agent’s Slack channel editor, choose a workspace if there is more than one and select Add to Slack. Everruns creates the app and opens Slack’s consent screen with that workspace selected. The channel then shows Live in Slack with an Open in Slack link.

See Connect a Slack workspace and Publish to Slack. PRs #3830, #3977.

Calls at xhigh and max reasoning effort on OpenAI and Azure run as background responses. If the stream drops or the worker restarts, Everruns picks the response back up instead of paying for it twice. A cancelled turn cancels the response, and the openai/background driver option overrides the policy.

See OpenAI. PRs #3932, #3969.

• An MCP client calls events/subscribe on /mcp to receive signed webhooks for session.completed, session.failed and session.input_required. Callbacks must be public HTTPS URLs and are verified before the first event.
• An agent trigger with trigger_type: "mcp_event" subscribes to an event one of the agent’s own MCP servers publishes, such as a tracker announcing new issues, and starts a session when it arrives.

See MCP and Agent triggers. PRs #3931, #3970.

The templates add GitHub review, issue upsert and pull request tools, enforced by the GitHub capability’s configuration, and a guided setup page.

See Set up review and security agents. PR #3980.

• GitHub pull request triggers and per-agent GitHub Apps: Agents connect their own GitHub App by clicks and act on pull requests, with PR triggers, a Connect GitHub card, PR tools, and a PR summarizer guide (#3935, #3943).
• OpenAI Responses hosted tools and remote MCP: The OpenAI Responses driver gains hosted web search (#3918, #3923), hosted code interpreter, shell and file search (#3928), remote MCP with approval and registered-server credentials (#3930), and structured output with background mode (#3932).
• Agent trigger event pipeline: A new event pipeline with filters, dedupe, and per-subject sessions (#3915), plus outbound MCP Events for session webhooks (#3931).
• MCP Apps for ChatGPT and Codex: Everruns serves MCP Apps views for ChatGPT and Codex (#3927), answering form-mode elicitation through ask_user (#3929).
• Pluggable decision drivers: Decision drivers route behind one router (#3921), with the OpenAI Decisions API available as a preview driver (#3924).
• Provider-neutral computer use: A provider-neutral computer tool runs on Browserless (#3922).
• New models: Claude Sonnet 5.5 joins as the favorite Sonnet (#3902) and GPT-6.1 Sol arrives with fast and ultrafast speed tiers (#3925); the sunset Claude Sonnet 4.5 profile is dropped (#3952).
• Rebuilt agent workspace: The agent page is rebuilt as a single view-and-edit workspace (#3917).

• Agent page rebuilt as one view-and-edit workspace
• Everruns MCP server renders as an MCP App in ChatGPT and Codex
• Claude Sonnet 5.5 and GPT-6.1 Sol with fast and ultrafast tiers
• GitHub pull request triggers with a per-agent GitHub App
• OpenAI hosted code interpreter, shell, file search and remote MCP

See the GitHub release and the full release notes for every pull request in this version.

{ "ref": "openai_server_tools",
  "config": { "tools": ["code_interpreter", "shell", "file_search"],
              "file_search_vector_store_ids": ["vs_abc123"] } }

Hosted calls show in the work log and are priced in usage. Each remote MCP call pauses on an approval card by default. A server that needs credentials is named by its registered Everruns MCP server, and its URL and token are resolved per call, never stored in agent config or events.

See OpenAI server tools. PRs #3918, #3923, #3928, #3930.

An agent can now act on GitHub pull requests. Use the Connect GitHub card on the agent’s Integrations tab to create and install the agent’s own GitHub App. The GitHub capability then gives it get_github_pull_request, get_github_pull_request_diff and upsert_github_comment.

Add a trigger in GitHub mode, or from the CLI with --github-events and --repositories, to start work on pull request events. Webhooks are signature verified, deduplicated by delivery ID, and routed to one session per pull request. Every trigger now shares that pipeline, so webhook triggers also accept filters and dedupe templates.

See Summarize GitHub pull requests. PRs #3935, #3943, #3915.

GPT-6.1 Sol joins the OpenAI models. Its controls.speed setting accepts fast and ultrafast. The OpenAI drivers reject a speed tier the model does not list before sending, and usage is priced at the tier that served the call.

Anthropic sunset Claude Sonnet 4.5, so its model profile and catalog entry are removed.

See Anthropic and OpenAI. PRs #3902, #3925, #3952.

The Everruns MCP server at /mcp is now an MCP App. In hosts that render MCP Apps, such as ChatGPT and Codex, running an agent shows a session view with status, messages and a reply box. Agent questions appear as a form, and pending approvals as Approve and Decline buttons that refuse stale clicks. An Everruns panel opens from the ChatGPT sidebar.

Hosts that do not render apps get the same tools as plain text, and the existing tool shapes are unchanged.

See MCP: In ChatGPT and Codex. PR #3927.

MCP servers, credentials, branding, starter files, network access, token usage and health check open as side sheets from More. Version history and archive move to the header’s More actions menu. Old /edit links open the page in edit mode, and /sessions opens the Sessions tab.

See Agent page. PR #3917.

• Per-organization Slack app provisioning: Each organization can provision its own Slack app rather than sharing one deployment-wide install (#3830), with concurrent installs for one endpoint serialized so a race can no longer create orphan apps or exhaust the config token (#3843).
• Agent app framework over Engine: An experimental framework for building agent apps directly on the Engine (#3802), with engine-level event listeners (#3804) and pluggable observability integrations (#3806).
• Server-side context compaction: Anthropic conversations compact context server-side, keeping long agent runs within budget (#3831).
• Ask User through tasks: The Ask User capability surfaces through the tasks channel, extending it beyond Platform Chat and MCP (#3828).
• Server hardening: Platform Chat question-answers and tool-results now require the session owner (#3842), unrequested reasoning summaries stay hidden from responses (#3840), feature-flag revocations take effect immediately (#3820), retained observer event memory is bounded (#3837), and native endpoint sessions are reused rather than rebuilt (#3814).

• Ask User questions reach MCP, A2A and Tasks clients
• Framework engines get event listeners and OpenTelemetry export

See the GitHub release and the full release notes for every pull request in this version.

• Over /mcp, the question set is a form-mode elicitation. A client that never declared elicitation gets the declared defaults.
• Over inbound A2A, the task reports that input is required, with the question as text and as a data part under everruns/ask_user. Reply with a data part under everruns/ask_user_answer.
• Over MCP Tasks, tasks/get exposes the pending form through inputRequests, and tasks/update answers it.

See Ask User, MCP and A2A. PRs #3781, #3780, #3828.

• Ask User over MCP and A2A: The Ask User capability now reaches agents beyond Platform Chat: it projects as form-mode elicitation over /mcp (#3781) and is carried over the inbound A2A channel as a DataPart (#3780).
• Always-thinking model reliability: Always-thinking models no longer return 400s or empty answers (#3782), preserved thinking keeps the conversation append-only (#3783), append-only clear_at reminders land (#3787), and hosted tool-search replay order is preserved (#3789).
• Server resilience under load: Background sweeps run isolated with their own database pool (#3786, #3779), and the server returns a graceful 503 on database pool exhaustion instead of failing hard (#3788).

• Claude always-thinking models no longer return 400s or empty answers
• Ask User lets agents pause and ask you questions mid-run

See the GitHub release and the full release notes for every pull request in this version.

In code: implement the AskUser trait to answer questions from your own interface when you embed the Framework.

See Ask User and Answer an agent’s questions. PRs #3735, #3738, #3747, #3757, #3790.

• Ask User: Agents can pause mid-run to ask a person a question: an inline UI card in Platform Chat (#3735), in-process and hosted strategies (#3738), expiry that resolves to declared defaults (#3757), and credential collection that never enters the log or agent context (#3747). Enabled on the default chat harnesses (#3737).
• Host Shell: A new host_shell capability runs the bash tool as real child processes on the machine hosting the agent, bounded by a kernel policy (Seatbelt and Landlock). Behind the host-shell feature on everruns or everruns-host; see the host shell example (#3719).
• Slack one-click install: Install the Slack app in one click, creating the app and taking its credentials straight from OAuth (#3724).
• New models: Claude Opus 5.5 lands as the recommended Opus (#3769), and GPT-6 Sol and Luna arrive with GPT-6 Luna as the default chat and utility model (#3771, #3774).
• Agent endpoint budget controls: Set spend budgets on agent endpoints from the UI (#3719).

• Agent endpoints and triggers replace Apps in Everruns
• Slack agents ask approval with buttons and show task progress
• Set spend budgets on agent endpoints from the Integrations tab
• Host Shell gives Framework agents a kernel-contained Bash
• MCP catalog presets, acts-as identity and personal connections
• Install an Everruns agent in Slack with one click
• Claude Opus 5.5, GPT-6 Sol and GPT-6 Luna models added

Before upgrading, read the breaking changes: Agent endpoints and triggers replace Apps in Everruns.

See the GitHub release and the full release notes for every pull request in this version.

• A request that ends on an assistant turn (prefill) fails fast with a configuration error that names the fix, instead of a provider 400 and a retry cycle.
• An explicit none reasoning effort maps to low on models where thinking cannot be turned off. Before, the driver sent nothing and the API ran at its own default.
• Each request in a session replays the previous one and only appends to it, so preserved thinking stays valid across turns.

See Anthropic. PRs #3782, #3783.

GPT-6 Luna replaces GPT-5.6 Terra as the platform default model. It is also the default utility model behind Agent Analyze, Health checks and llm_judge guardrails; set UTILITY_LLM_MODEL to choose another.

See Anthropic, OpenAI and environment variables. PRs #3769, #3771, #3774.

One-click install needs a Slack app configuration token on the deployment, which Everruns Cloud has. Without one, the button is hidden and the manual fields stay available under Configure manually.

See Publish to Slack and Slack. #3724

In agent config: an MCP entry can reference a preset with catalog:<name> and say whose sign-in its calls use with actsAs (none, service, user or user_or_service):

{ "github": { "use": "catalog:github", "actsAs": "user" } }

See MCP and MCP credentials. #3726, #3675, #3638

In the UI: with the app_budgets feature flag on, the agent’s Integrations tab manages agent and endpoint budgets, showing the period, its UTC reset and any refusal. Members without budget.manage see them read-only.

Slack and FCP endpoint budgets now bind and are enforced. If you configured one before and it never took effect, it starts limiting spend.

See Budgets and Enforce a budget. #3719, #3623, #3625

• Approvals: when an agent pauses with request_approval, the thread shows buttons. Only the person the agent is answering can use them, and answering removes the buttons. Apps created earlier must save a fresh manifest once.
• Task progress: a delegating agent posts one message, updated in place, that summarizes its sub-tasks, for example how many finished or failed.
• Slack actions: a native slack capability adds slack_add_reaction, slack_update_message, slack_lookup_user and slack_upload_file, using the endpoint’s own bot token instead of a separate Slack MCP server.

See Slack approvals, task progress and Slack actions. #3696, #3697, #3693

• Platform Chat v2 on a unified command line: One command line serves both people and agents, with Platform Chat rebuilt on top of it (#3680).
• Soft approval: A soft-approval capability lands and is on by default for Generic and Platform Chat, so an agent can proceed with a lightweight, reversible approval step (#3655).
• Typed judgments and a guardrail engine: Typed judgments become a capability, backing a guardrail engine for classifying and gating agent behavior (#3644).
• Scoped MCP catalogs and acts-as: MCP gains scoped catalog presets and acts-as credential semantics, with endpoint auth stored separately (#3638, #3635).

• Framework direct model calls with Model::complete
• One everruns command line for the CLI, MCP and agent shells
• Guardrails can use typed judgments as their check engine
• Soft approval: agents pause before destructive or outward actions

See the GitHub release and the full release notes for every pull request in this version.

In the UI: an agent’s Integrations tab lists its endpoints and triggers, each with its own publish switch and real endpoint URL, plus a switch to suspend all of the agent’s exposures at once. A new Exposures page lists every endpoint and trigger in the organization, flags anonymous live ones, and can suspend or resume an agent. App management pages are gone.

In the API: publishing moves one endpoint. Suspend with POST /v1/agents/{agent_id}/exposures/suspend and resume with .../resume. App write and command surfaces are removed; existing App records are frozen and ingress URLs keep working.

See Channels and Agent triggers. #3698, #3624, #3627, #3621

It is enabled at normal on the Generic and Platform Chat harnesses. The approver is resolved on the server from the authenticated message, never from what the model says, and recorded as agent.approval.requested and agent.approval.granted in the audit log.

See Soft approval. #3655

Operators enable it with the UTILITY_TYPESAFE_API_KEY deployment credential. When it is unset, jev checks are skipped with a log line and the turn proceeds. An experimental jev capability with a jev_evaluate tool is available in development registries.

See Guardrails and TypeSafe. #3644

everruns agents list
everruns sessions --help

In code: a Framework application exposing its own command tree can name its root token with CliCommandSource::root instead of everruns.

A Platform Chat v2 harness built on this command line is available behind an organization opt-in flag. See CLI and CLI reference. #3680, #3490, #3496

Before this change, a machine set to dark mode got a partly dark interface; the whole UI now switches together, with semantic status colors that work in both themes.

See UI. #3637, #3639

• Agent exposure management: An agent’s endpoints and triggers now have one home, and a new cross-agent Exposures view lists exposure across all agents. Publish moves to the endpoint, and an agent’s exposure can be suspended (#3624, #3627, #3621).
• Endpoint-attributed sessions and budgets: Sessions and budgets now attribute to the endpoint they arrive through, and Slack and FCP endpoint budgets bind and enforce their limits. Previously configured endpoint budgets that were inert will begin enforcing (#3623, #3625).
• Channels re-parent from App to Agent: Channels move from the retired App abstraction to the agent, with grandfathered apps synthesized into agents automatically (#3618, #3598).

• The Everruns UI supports dark mode with a theme setting

See the GitHub release and the full release notes for every pull request in this version.

Set agent_surface_enabled on a Slack channel and its app also appears as an agent in Slack’s dedicated assistant pane, while @mentions in channels keep working. In the pane, replies stream as the agent writes them, the stop button cancels the running turn, and the agent’s status and thread title show while a turn runs. Thread context is persisted and given to the agent, and replies post as markdown blocks.

The generated manifest declares the agent feature and fills Slack’s suggested prompts from the agent’s conversation starters, or the harness’s. An existing app must update its manifest and reinstall to gain the pane.

See Slack and Publish to Slack. #3560, #3564, #3606

• Slack agent surface: Agents now run natively in Slack. The agent surface is enabled end to end: replies stream token-by-token into the pane, a stop button cancels the running turn, agent status and the thread title show while a turn runs, thread context is persisted and fed to the agent, and replies post as markdown blocks with correlation metadata (#3560, #3564, #3568, #3581, #3567, #3563).
• Session environments: A session now declares where its commands run and what they may touch, giving agents an explicit execution environment (#3599).
• Endpoint-ID ingress routing: Ingress routes mount on endpoint IDs, with reserved routing tag prefixes to keep session routing unambiguous (#3593, #3589).
• Agent-owned exposure replaces Apps: The App abstraction is retired from the product surface in favor of agent-owned exposure (#3590, #3585).

• Sessions show where their commands run and what they may touch
• Everruns agents run in Slack’s agent pane

See the GitHub release and the full release notes for every pull request in this version.

In the UI: the session Workspace tab gains an Environment panel.

In the API: GET /v1/sessions/{id}/environment and GET /v1/environment-targets are read-only views of the same data. The feature is enabled wherever sandboxes are.

In code: the Framework’s Environment gains named compute and containment members, with HostCompute behind the host-compute feature.

See Workspaces and environments. #3599

• PDF input with file attachments: Agents can now accept PDF documents as file attachments alongside text input (#3535).
• Muse image model: A new Muse image-generation model with a configurable fallback (#3523).
• Quick-connect provider grid: Connecting an LLM provider is now a one-click grid in the UI (#3526).
• Native async tool execution on durable workers: Long-running tool calls execute natively on durable workers instead of blocking a turn (#3466).
• Synchronous wait-for-completion: A non-streaming wait-for-completion path spanning drivers, the sessions API, and the framework (#3541).
• Agent discovery documents: The server serves agent discovery documents, and create_agent/session/message descriptions surface their argument constraints (#3533, #3530).

• Everruns servers publish an MCP server card and auth.md
• Connect an LLM provider with one pasted key, checked before saving
• Image generation adds the Muse image model with a fallback
• Session schedules no longer fire twice on multi-replica deployments
• Attach PDF files to agent messages
• Send a message and wait for the turn result in one request

See the GitHub release and the full release notes for every pull request in this version.

In the API: POST /v1/sessions/{id}/messages?wait=true returns 200 with the turn result when the turn completes or fails, or 202 with partial messages when timeout_ms (default 120 s, at most 600 s) expires. Without wait it still returns 201 immediately, so existing clients are unaffected.

In code: RunOptions::timeout bounds Session::run_with; when it expires the in-process turn is cancelled and its outcome returned.

See Sessions and Stream events. #3541

In the UI: the chat composer offers PDF attachments when the selected model accepts PDF input, and shows each attachment as a file chip.

In the API: upload a PDF with POST /v1/files (PDF only, up to 32 MB), then reference it as a file content part in the message. GET /v1/files, GET /v1/files/{id} and DELETE manage stored files. Everruns converts the attachment for each provider, including Anthropic documents, Gemini inline data, Bedrock documents and OpenAI file inputs.

See UI. #3535

See Session schedules. #3527

platform_management is retired. Its 14 hand-written tools are replaced by the platform capability, which derives the same surface from the command catalog. Agents that still list platform_management should switch to platform.

See Platform Management (removed) and Platform. #3513

Platform Chat also runs on a wider capability set, adding current time, message timestamps, a to-do list, prompt caching, tool call repair and parallel tool calls. Existing organizations pick these up without a migration.

See Platform Chat. #3508, #3507

{ "config": { "fallback": "auto" } }

Credentials now come only from configured providers. The capability no longer reads OPENAI_API_KEY or OPENAI_BASE_URL from session secrets.

See OpenAI image generation. #3523

In the API and CLI: POST /v1/providers/check-credentials checks a key without storing it:

everruns providers check-credentials check --provider-type openai --api-key "$OPENAI_API_KEY"

Disabled models can be re-enabled. See Providers and CLI providers. #3526, #3516, #3515

• Hierarchical AGENTS.md as conversation context: Agent instructions now resolve AGENTS.md files hierarchically and are injected as the leading user-role message of each turn, keeping untrusted workspace instructions below harness safety instructions in the instruction hierarchy (#3501).
• Platform Chat for every user: Each user gets a precreated, pinned Platform Chat thread, and the Platform Chat capability set is widened so it can drive more of the control plane (#3508, #3507).
• Capability deprecation lifecycle: Capabilities can now be marked Deprecated (still functional, surfaces warn) or Retired (inert and hidden), and the platform_management capability was retired via the new lifecycle (#3513).
• Unified everruns command tree for MCP and CLI: A noun-verb command tree is now exposed over MCP, and a host can name the root token of its own command tree from the CLI (#3490, #3496).
• Safer provider onboarding and model management: Provider API keys are validated before being stored, a configured key reliably produces a usable org, discovered models are adopted instead of conflicting on explicit create, and disabled models can be re-enabled (#3516, #3515, #3517, #3512).

• AGENTS.md files resolve hierarchically as conversation context
• Every Everruns user starts in a pinned Platform Chat thread
• platform_management capability retired in favor of platform

Before upgrading, read the breaking changes: platform_management capability retired in favor of platform.

See the GitHub release and the full release notes for every pull request in this version.

The descriptions of create_agent, create_session and create_message also state their argument constraints, so an agent gets them right on the first call.

See Authentication. #3533, #3530

This keeps untrusted repository instructions below harness safety instructions and keeps the system prompt stable for caching. Resolution is bounded to 16 files, 32 KiB each and 128 KiB in total.

See Agent instructions and Use AGENTS.md. #3501

• ~3.5x faster agent startup: Agent startup-to-ready latency cut by roughly 3.5x (#3486).
• MCP URL-mode elicitation with a consent pause: MCP elicitation now works in URL mode on both sides, with an explicit consent pause surfaced in the UI (#3354).
• Gen-AI and OpenInference trace conventions: Observability now emits Gen-AI agent and OpenInference trace conventions for standard agent tracing (#3353).
• Expanded model catalog: Added OpenAI GPT-6 Astra, Claude Fable 5.1, and Muse Spark 1.3 profiles with live-matrix coverage (#3363, #3351, #3352).
• Broad provider wire-contract hardening: A large sweep of driver and provider fixes plus expanded contract tests across Anthropic, OpenAI, Gemini, OpenRouter, Bedrock, MAI, Meta, and Fireworks, tightening streaming, usage, reasoning, and request-shaping behavior.

• First message of a new chat no longer drops or shows out of order
• Everruns CLI plugins, skills and knowledge-bases commands work again
• Organization default model can fall back to the platform default
• Only organization admins can rename an organization
• Tool calls are validated against their schema before running
• OpenTelemetry traces follow Gen-AI and OpenInference conventions
• MCP URL elicitation pauses for consent in the Everruns UI
• GPT-6 Astra, Claude Fable 5.1 and Muse Spark 1.3 model profiles
• Framework agents start about 3.5 times faster

See the GitHub release and the full release notes for every pull request in this version.

• OpenAI GPT-6 Astra (gpt-6-astra), seeded as an enabled favorite. Its reasoning effort accepts a new max tier above xhigh.
• Claude Fable 5.1 (claude-fable-5-1 and its 1M-context variant), seeded as enabled favorites with adaptive thinking.
• Muse Spark 1.3 and its Contributor tier on the Meta Model API. Muse Spark 1.2 stays selectable.

New organizations get them on the Models page automatically.

See Providers, Meta and Anthropic. #3363, #3351, #3352

In the UI: a tool call that stops on a URL elicitation pauses the turn on a confirmation card. The consent is single use and bound to the server, tool and domain you were shown.

For MCP clients of Everruns: the /mcp endpoint gains session_set_secret and connect, which point you at a form Everruns serves instead of taking the value.

See MCP URL elicitation and Complete a URL elicitation. #3354

Choose the vocabularies with EVERRUNS_TRACE_CONVENTIONS (default gen_ai,openinference):

# OpenInference only, for example for Arize Phoenix
export EVERRUNS_TRACE_CONVENTIONS=openinference

See OpenTelemetry and Framework observability. #3353

For OpenAI Chat Completions and Open Responses, compatible tools are now sent with strict schemas. A schema that cannot be converted safely is sent non-strict, as before.

See Tool Call Repair and #3336.

Admins see no change. See #3315.

To go back to the platform default after choosing a model, select the platform default option, or send "default_model_id": null in PATCH /v1/orgs/{org}. Omitting the field leaves the setting unchanged.

See CLI: models, #3114 and #3321.

The chat composer is also locked when no explicit or default model resolves, with the model menu left enabled to fix it.

See Stream events, #3296 and #3295.

Replay to providers is also more accurate: Anthropic thinking blocks keep their own signatures, including redacted_thinking; OpenAI reasoning items keep the ids the provider issued; and the Gemini driver now requests, parses and replays thoughts.

See Event reference, #3276 and #3279.

Highlights

• Reasoning modeled as ordered artifacts: Model reasoning is now captured as ordered artifacts with a published classification, so reasoning steps replay in the order they were produced and their kind is visible to consumers (#3276).
• Bounded batch file reads: The filesystem capability gained bounded batch reads, letting agents fetch several files in a single guarded call instead of one round-trip per file (#3275).

• Agents read several workspace files in one read_many_files call
• Global chat session endpoints removed from the Everruns API
• Messages API reports reasoning in order and the message phase

Before upgrading, read the breaking changes: Global chat session endpoints removed from the Everruns API.

See the GitHub release and the full release notes for every pull request in this version.

After upgrading, POST /v1/sessions/chat returns 405 and POST /v1/sessions/chat/voice returns 404. Sessions already tagged global-chat stay listed and readable through the normal session routes, and the Platform Chat harness is unchanged.

See Platform Chat and #3278.

Models that emit one tool call per response no longer need one model round per file. In the PR’s tests, three independent reads took 2 model calls instead of 4 and about half the input tokens. Loop detection recognizes repeated batch reads.

See File System and #3275.

The Work, Events and Workspace tabs now carry counts. Empty tabs show no badge, and large counts are shortened, for example 12.4K.

See Web UI, Environment variables, #3260 and #3266.

Highlights

• Chat with harnesses, archive chats, and generated run summaries: Sessions gained interactive chats against harnesses, the ability to archive chats and toggle showing all of them, and automatically generated run summaries, with count badges on session detail tabs (#3262, #3265, #3260, #3266).
• Connection-level custom provider headers: Providers can now carry custom headers at the connection level, with cache diagnostics to see how they resolve; header values are redacted in diagnostics (#3232, #3240).
• Leaner supply chain and smaller binaries: Owning 114 small third-party contracts in-tree removed those dependencies from the build, and stripping plus CLI dispatch boxing shrank shipped binaries by ~22% (#3245, #3227).

• Everruns CLI, server and worker binaries are about 20% smaller
• Start a chat with any harness, not only an agent
• Provider connections carry custom headers and cache diagnostics
• Archive finished chats and show them again on demand
• Session pages show a generated run summary and tab counts

See the GitHub release and the full release notes for every pull request in this version.

In the UI: use the archive control on a row in Chats, and the Show archived filter to list archived threads. The sidebar’s All chats link stays visible even when no threads are listed.

In the API: PUT /v1/sessions/{id}/archive archives and DELETE /v1/sessions/{id}/archive restores. Both are idempotent. Session lists hide archived sessions unless you pass include_archived=true, which widens facet counts too.

See Web UI and #3265.

In the UI: open a provider in Settings → Providers and use the Advanced card to add Custom headers or turn on Prompt cache diagnostics.

In the API: the provider settings accept the same options. Transport headers such as host and content-length are rejected, the list is limited to 16 headers of up to 2 KB each, and header values are redacted in diagnostics. The Anthropic, Gemini, Chat Completions and Open Responses drivers apply them.

See Providers, #3232 and #3240.

The thread names the harness as its counterpart and appears in the Chats list like any other chat.

See Harnesses, Generic harness and #3262.

The CLI binary went from 18.1 MB to 14.0 MB (22.4%). The server and worker binaries are 18.3% and 19.0% smaller from stripping alone. Behavior is unchanged.

See CLI and #3227.

Highlights

• Register capabilities on a running runtime: A host holding Arc<InProcessRuntime> can now make a capability discovered after composition resolvable without rebuilding the runtime, via register_capability/is_capability_registered. Registration stays separate from activation, and readers always see a consistent registry snapshot (#3224).
• A2A delegation is now opt-in: Outbound A2A delegation moved behind a new a2a Cargo feature, off by default, so the standard build no longer drags a second HTTP/TLS stack into consumers that only run the local runtime. Hosts that delegate to remote A2A agents enable the a2a feature explicitly (#3221).

• A2A delegation in the everruns crate is now an opt-in feature
• Register a capability on a running Framework runtime

Before upgrading, read the breaking changes: A2A delegation in the everruns crate is now an opt-in feature.

See the GitHub release and the full release notes for every pull request in this version.

Highlights

• First-class sandbox and checkpoint records: Sandboxes and checkpoints are now first-class records, giving durable runs an explicit, addressable execution substrate (#3197).

See the GitHub release and the full release notes for every pull request in this version.

Highlights

• The 0.18 neutral-kernel boundary: everruns-core is now a neutral execution kernel. Hosted platform records, sessions, agents and agent versions, harnesses, evals and observers, connectors, system email, feature flags, and OAuth/credential infrastructure, moved out to everruns-platform (and sibling crates), and the logic-free everruns-runtime compatibility crate was removed. REST/gRPC shapes and stored schema are unchanged; see the 0.18 migration guide and the Breaking notes below (#3160, #3101).
• Engine-owned sessions and unified execution: Engines now own the session lifecycle, and a single execution path serves both in-process and durable runs (#3179, #3173, #3167).
• Multi-head workspaces: Framework workspaces now support multiple heads (#3166).
• Metered embedding spend: Retrieval and knowledge-index embedding spend is now metered and ledgered against the organization (#3149, #3163).
• Chats as core functionality: Chats is always present in navigation and search for every organization with no feature opt-in, and now supports chat pinning.

• Chats is on for every organization, with pinned chats
• Session recordings split into Transcript, Timeline and Events
• Knowledge index embedding spend is metered and ledgered
• Framework workspaces gain multiple isolated Git heads
• everruns-core 0.18 becomes a neutral execution kernel

Before upgrading, read the breaking changes: everruns-core 0.18 becomes a neutral execution kernel.

See the GitHub release and the full release notes for every pull request in this version.

Knowledge index sync, which embeds every indexed chunk in the background, now writes its spend to the usage ledger against the organization. Sync spend is not debited from any user’s budget, because no session triggered it.

See Knowledge Index, Budgets, #3149 and #3163.

• Transcript is the default view: a full-width, read-only replay of the conversation.
• Timeline is the execution history, with noisy worker details collapsed.
• Events is the complete raw event ledger.

Session links, global search and Fork into chat open the Transcript. Work, Workspace and Cost behave as before.

See Web UI and #3113.

Highlights

• Live, steerable sessions: Sessions are now live and steerable, with canonical session events, session history and resume, and a session work and wake API (#3095, #3090, #3089, #3083).
• Sessions as the operational list: The Sessions surface was rebuilt as the operational list, recording session source and offering a filterable list with facet counts (#3078, #3073).
• Advanced capability authoring: New advanced capability authoring API with typed lifecycle hooks, a workspace access policy, unified capability configuration, and separated provider and model configuration (#3084, #3086, #3085, #3096, #3093).
• Daytona workspace recovery: Sandboxes recover Daytona workspaces after instance loss (#3092).

See the GitHub release and the full release notes for every pull request in this version.

Chat threads can also be pinned: use the pin button in a thread header to keep it at the top of the Chats list and the sidebar.

See Web UI, Platform Chat and #3100.