Everruns Framework changelog
October 2026
serve keeps daemon data in S3 and can require API keys
serve start --store s3:// keeps a daemon's data in an object store, agent routes can require API keys or OIDC, and serve resumes cut-off sessions at boot.
Explicit communication lets agents choose when to reply
An agent's Communication setting can be Explicit, so assistant text stays as private notes and the agent speaks only through send_message or stays silent.
Agent budgets stop turns before each model call
Agent budgets are now checked before every model call in the native agent loop, so an exhausted budget stops the turn with a typed reason and budget events.
Voice channels let people call an agent
A voice channel lets people talk to an agent over WebRTC while it keeps its own model, tools and instructions, on the platform and in the Framework.
Everruns 0.47.0: Codex session reliability
Everruns 0.47.0 was released on 2026-10-09 with 1 highlight and 5 changelog entries. The release notes and any upgrade notes are linked here.
Everruns 0.46.0: Agents home
Everruns 0.46.0 was released on 2026-10-09 with 2 highlights and 1 changelog entry. The release notes and any upgrade notes are linked here.
Everruns 0.45.0: Tools in Shell
Everruns 0.45.0 was released on 2026-10-09 with 3 highlights and 2 changelog entries. The release notes and any upgrade notes are linked here.
Fix ChatGPT plan sessions failing with HTTP 400
Sessions that use a ChatGPT plan connection no longer fail with HTTP 400, because Everruns stops sending request-body metadata the Codex backend rejects.
Everruns 0.44.0: A2A 1.0 HTTP+JSON and PACT
Everruns 0.44.0 was released on 2026-10-08 with 4 highlights and 4 changelog entries. The release notes and any upgrade notes are linked here.
Coordinator threads in Platform Chat and the Framework
Platform Chat and Framework agents can hand focused work to child threads and keep a board of what is working, what needs you and what is ready for review.
Everruns 0.43.0 release
Everruns 0.43.0 was released on 2026-10-07 as a maintenance release with fixes and internal changes. The full release notes are linked here.
Claude Haiku 5.5 model added for Anthropic providers
The claude-haiku-5-5 model and its 1M-context twin are seeded as an enabled favorite, with adaptive thinking, tool search and the computer toolset.
Everruns 0.42.0 release
Everruns 0.42.0 was released on 2026-10-06 with 3 changelog entries. The release notes and any upgrade notes are linked here.
Framework crates consolidated into everruns-core and the facade
Engine, host, builtins, MCP, AG-UI, driver and contract crates merged into fewer packages, so Framework users update dependencies and imports when upgrading.
Everruns 0.41.0 release
Everruns 0.41.0 was released on 2026-10-04 with 1 changelog entry. The release notes and any upgrade notes are linked here.
Everruns 0.40.0 release
Everruns 0.40.0 was released on 2026-10-04 as a maintenance release with fixes and internal changes. The full release notes are linked here.
Everruns 0.39.0 release
Everruns 0.39.0 was released on 2026-10-04 with 2 changelog entries. The release notes and any upgrade notes are linked here.
Everruns 0.38.0: Personal ChatGPT plan connections
Everruns 0.38.0 was released on 2026-10-03 with 2 highlights and 4 changelog entries. The release notes and any upgrade notes are linked here.
Everruns 0.37.0: Approval history in context
Everruns 0.37.0 was released on 2026-10-03 with 4 highlights and 1 changelog entry. The release notes and any upgrade notes are linked here.
Everruns 0.36.0: Choose an Agent Environment
Everruns 0.36.0 was released on 2026-10-03 with 3 highlights. The release notes and any upgrade notes and every pull request are linked here.
Everruns 0.35.0: Safer sandbox tools
Everruns 0.35.0 was released on 2026-10-03 with 3 highlights and 1 changelog entry. The release notes and any upgrade notes are linked here.
Conversation becomes the default harness and Generic is deprecated
Built-in harnesses form progressive presets from Base to Worker, Conversation is the new default, and agents that inherited Generic are pinned to it on upgrade.
Host Everruns serve apps on Amazon Bedrock AgentCore Runtime
The experimental everruns-serve-agentcore crate runs a serve app on AgentCore Runtime with AG-UI streaming, a WebSocket transport and Bedrock models.
Run an Everruns serve app durably on celld
The everruns-serve-celld hosting target snapshots and restores a serve app's state, so sessions survive losing the serve process or the celld node mid-turn.
everruns-durable workflow engine published on crates.io
The durable workflow engine behind Everruns ships as its own crate with timers, child workflows and a bundled idempotent PostgreSQL schema you can migrate.
Everruns 0.34.2: Execution kernel in a JavaScript isolate
Everruns 0.34.2 was released on 2026-10-02 with 3 highlights and 2 changelog entries. The release notes and any upgrade notes are linked here.
Everruns 0.34.1 release
Everruns 0.34.1 was released on 2026-10-02 with 2 changelog entries. The release notes and any upgrade notes are linked here.
Everruns 0.34.0: AG-UI 1.0
Everruns 0.34.0 was released on 2026-10-02 with 10 highlights and 6 changelog entries. The release notes and any upgrade notes are linked here.
Cloudflare AI Gateway and Vercel AI Gateway model providers
Agents can run models through Cloudflare AI Gateway or Vercel AI Gateway, and the Cloudflare provider lists Workers AI text-generation models automatically.
AG-UI 1.0 endpoint and AG-UI streaming for Framework sessions
The AG-UI channel moves to protocol 1.0 with interrupts, resume, frontend tools and subagents, and any Framework session can serve AG-UI from your own server.
A2A 1.0 endpoint with task results, streaming and push notifications
Everruns A2A channels negotiate protocol 1.0, return agent output as task artifacts, support ListTasks and SubscribeToTask, and send push notifications.
OpenAI structured output and background mode for long reasoning
OpenAI drivers enforce JSON Schema structured output, and xhigh or max reasoning calls run as background responses that resume after a drop or worker restart.
Everruns 0.33.0 release
Everruns 0.33.0 was released on 2026-10-01 with 8 highlights and 5 changelog entries. The release notes and any upgrade notes are linked here.
September 2026
Claude Sonnet 5.5 and GPT-6.1 Sol with fast and ultrafast tiers
Claude Sonnet 5.5 becomes the favorite Sonnet model, GPT-6.1 Sol adds fast and ultrafast speed tiers, and the sunset Claude Sonnet 4.5 profile is removed.
Everruns 0.32.0: Per-organization Slack app provisioning
Everruns 0.32.0 was released on 2026-09-26 with 5 highlights and 2 changelog entries. The release notes and any upgrade notes are linked here.
Framework engines get event listeners and OpenTelemetry export
Register a listener once per Engine to receive events from every session it owns, and export Framework sessions to OpenTelemetry or Braintrust.
Everruns 0.31.0: Ask User over MCP and A2A
Everruns 0.31.0 was released on 2026-09-24 with 3 highlights and 2 changelog entries. The release notes and any upgrade notes are linked here.
Ask User lets agents pause and ask you questions mid-run
Agents with the ask_user capability pause a run to ask one to four choice, text or secret questions, shown as an inline card that you answer in place.
Everruns 0.30.0: Ask User
Everruns 0.30.0 was released on 2026-09-23 with 5 highlights and 7 changelog entries. The release notes and any upgrade notes are linked here.
Claude always-thinking models no longer return 400s or empty answers
The Anthropic driver rejects prefill early, maps an explicit none effort to low, and keeps requests append-only on Opus 5.5 and other always-thinking models.
Host Shell gives Framework agents a kernel-contained Bash
The new host_shell capability lets an embedded Everruns agent run real host processes inside a Seatbelt or Landlock policy that fails closed.
Everruns 0.29.0: Platform Chat v2 on a unified command line
Everruns 0.29.0 was released on 2026-09-18 with 4 highlights and 4 changelog entries. The release notes and any upgrade notes are linked here.
Soft approval: agents pause before destructive or outward actions
The new soft_approval capability, on by default for Generic and Platform Chat, has agents ask before risky actions and records who approved in the audit log.
One everruns command line for the CLI, MCP and agent shells
Every Everruns operation now has one noun-verb spelling with --help that works the same in the everruns CLI, the /mcp endpoint and an agent's shell.
Framework direct model calls with Model::complete
The Everruns Framework adds a promoted path for one prompt and one answer, and each driver now reads its vendor's standard credential environment variables.
Everruns 0.28.0: Agent exposure management
Everruns 0.28.0 was released on 2026-09-16 with 3 highlights and 1 changelog entry. The release notes and any upgrade notes are linked here.
Everruns 0.27.0: Slack agent surface
Everruns 0.27.0 was released on 2026-09-15 with 4 highlights and 2 changelog entries. The release notes and any upgrade notes are linked here.
Sessions show where their commands run and what they may touch
Each session now declares its execution environment, readable through the API and shown in an Environment panel on the session Workspace tab.
Everruns 0.26.0: PDF input with file attachments
Everruns 0.26.0 was released on 2026-09-13 with 6 highlights and 6 changelog entries. The release notes and any upgrade notes are linked here.
Send a message and wait for the turn result in one request
The sessions API can now hold a message request open until the turn finishes, and Framework runs accept a timeout, for callers that do not consume SSE.
platform_management capability retired in favor of platform
Capabilities gain a deprecated and retired lifecycle, and platform_management is the first one retired: it stays registered but no longer contributes tools.
Everruns 0.25.0: Hierarchical AGENTS.md as conversation context
Everruns 0.25.0 was released on 2026-09-11 with 5 highlights and 3 changelog entries. The release notes and any upgrade notes are linked here.
AGENTS.md files resolve hierarchically as conversation context
Agent instructions now load AGENTS.md from the workspace root down to the working directory and place them below harness safety instructions.
Everruns 0.24.0: ~3.5x faster agent startup
Everruns 0.24.0 was released on 2026-09-09 with 5 highlights and 9 changelog entries. The release notes and any upgrade notes are linked here.
Framework agents start about 3.5 times faster
Building an agent and creating a session in the Everruns Framework now takes roughly a third of the time, with no change to steady-state behavior.
GPT-6 Astra, Claude Fable 5.1 and Muse Spark 1.3 model profiles
Everruns adds model profiles for OpenAI GPT-6 Astra, Claude Fable 5.1 and Muse Spark 1.3, seeds them for new organizations and covers them in live tests.
OpenTelemetry traces follow Gen-AI and OpenInference conventions
Agent turn traces now carry the OpenTelemetry Gen-AI agent attributes and OpenInference attributes on the same spans, selectable per deployment.
August 2026
Tool calls are validated against their schema before running
Everruns checks every tool call's arguments against the tool's JSON Schema and enables strict schemas for compatible OpenAI Chat and Responses tools.
Messages API reports reasoning in order and the message phase
Session messages now carry ordered reasoning parts plus phase and phase_source, so clients can tell intermediate commentary from the final answer.
Everruns 0.22.0: Reasoning modeled as ordered artifacts
Everruns 0.22.0 was released on 2026-08-25 with 2 highlights and 3 changelog entries. The release notes and any upgrade notes are linked here.
Agents read several workspace files in one read_many_files call
The File System capability adds read_many_files, which reads 2 to 10 independent files in one ordered call and halves model round trips for discovery.
Everruns 0.21.0 release
Everruns 0.21.0 was released on 2026-08-22 with 3 highlights and 5 changelog entries. The release notes and any upgrade notes are linked here.
Register a capability on a running Framework runtime
A host holding a shared InProcessRuntime can add a capability discovered after startup and activate it in live sessions without rebuilding the runtime.
Everruns 0.20.0: Register capabilities on a running runtime
Everruns 0.20.0 was released on 2026-08-19 with 2 highlights and 2 changelog entries. The release notes and any upgrade notes are linked here.
Everruns 0.19.0: First-class sandbox and checkpoint records
Everruns 0.19.0 was released on 2026-08-18 with 1 highlight. The release notes and any upgrade notes and every pull request are linked here.
A2A delegation in the everruns crate is now an opt-in feature
Outbound A2A delegation moved behind a new a2a Cargo feature that is off by default, so local runtime builds no longer compile a second HTTP and TLS stack.
Everruns 0.18.0: The 0.18 neutral-kernel boundary
Everruns 0.18.0 was released on 2026-08-16 with 5 highlights and 5 changelog entries. The release notes and any upgrade notes are linked here.
everruns-core 0.18 becomes a neutral execution kernel
Hosted platform records and services moved out of everruns-core into everruns-platform and sibling crates, and the everruns-runtime crate was removed.
Framework workspaces gain multiple isolated Git heads
Framework sessions can bind to one named head of a workspace, such as a local Git worktree, with fork, archive and destroy as explicit operations.
Everruns 0.17.26: Live, steerable sessions
Everruns 0.17.26 was released on 2026-08-10 with 4 highlights. The release notes and any upgrade notes and every pull request are linked here.