Release history
agency-swarm releases
Reliable Multi-Agent Orchestration Framework
All releases
13 shown
Fixed Python 3.14 startup compatibility.
Full changelog
Fix Python 3.14 startup compatibility by updating datamodel-code-generator and adding Python 3.14 coverage.
- Sanitize legacy unencrypted reasoning spans before browser‑auth replay to prevent leakage
Full changelog
Patch release for Agent Swarm TUI / browser-auth history replay.
Changes:
- Fix stateless OpenAI Responses reasoning replay when browser-auth runs with store=false.
- Request reasoning.encrypted_content for store=false replay paths.
- Sanitize legacy unencrypted reasoning spans before replay.
- Preserve per-run ModelSettings boundaries and avoid leaking resolved agent settings into reusable RunConfig objects.
- Add focused coverage plus a live OpenAI stateless reasoning replay proof.
- Manual chat_history clients must replay required inline attachment parts unless they use future server‑managed state or durable file IDs.
- Accept structured Responses‑style message attachments in FastAPI requests
Full changelog
Patch release for structured FastAPI message attachments used by AgentSwarm TUI file attachment flows.
- Accept structured Responses-style message attachments in FastAPI requests.
- Preserve streamed Codex output ordering and function-call deduplication.
- Document that manual chat_history clients must replay required inline attachment parts unless they use future server-managed state or durable file IDs.
Minor fixes and improvements.
Full changelog
What's Changed
- fix(fastapi): warn-and-drop client_config.litellm_keys when litellm missing by @bonk1t in https://github.com/VRSEN/agency-swarm/pull/636
Full Changelog: https://github.com/VRSEN/agency-swarm/compare/v1.9.4...v1.9.5
Minor fixes and improvements.
Full changelog
Changes since v1.9.3
- chore(cli): bump _CLI_VERSION to 1.4.24.
- fix: missing event handling for codex api (#633).
Fixed FastAPI Codex client_config.base_url scoping for non-OpenAI-compatible LiteLLM providers.
Full changelog
Bugfixes
- fix(fastapi): scope Codex client_config.base_url by LiteLLM provider (#630) — closes #629
This patch preserves the documented client_config.base_url + litellm_keys proxy pattern, only strips the Codex OAuth base_url when the target provider is non-OpenAI-compatible, and adds focused tests around the provider-scoped behavior.
- Per-request agent model override via client_config.model in FastAPI requests
Full changelog
Bugfixes
- Contain Agent Swarm CLI TUI bridge output to the bridge thread and its async children so other threads' stdout/stderr/logging no longer get silenced (#618).
Features
- Per-request agent model override via
client_config.modelin FastAPI requests. Callers can swap an agent's model for a single request while preserving the agent's existing transport (OpenAI client or LiteLLM base_url/api_key), GPT-5 model-family defaults, and OpenClaw usage aliases (#621).
Fixes browser-auth Codex backend error, Python bridge output containment in TUI, and session memory recovery.
Full changelog
Bugfixes
- Fix browser-auth "Store must be set to false" 400 from the Codex Responses backend when generating chat titles and when request-scoped client overrides retarget agents to Codex (#612).
- Contain Python bridge stdout/stderr writes and existing stream-handler logger output so they no longer leak into the Agent Swarm CLI TUI; replaces the previous print()-only trap with stream-level containment (#613).
- Pin Agent Swarm CLI to 1.4.12, which recovers local Agency Swarm session memory across loopback port changes (#64 in agentswarm-cli).
- FastAPI wildcard CORS origins no longer automatically enable credentials
- Metadata response shows absolute paths instead of relative allowed_local_file_dirs
- New TUI does not yet support tui(show_reasoning=False) parameter
- npx @vrsen/agentswarm launches TUI with automatic project setup
- OpenClaw runtime, proxy, and agent support built-in
- Request-scoped FastAPI control for OpenAI and LiteLLM models
Terminal UI auth and onboarding fixes from agentswarm-cli 1.4.7.
- raise_input_guardrail_error is canonical; throw_input_guardrail_error deprecated; return_input_guardrail_errors removed
- Handoff now maps to Agency Swarm handoff tool; OpenAI SDK handoff as SDKHandoff
- Built-in PresentFiles tool for file preview management
- Conversation starters and quick replies cache improvements
- Tool input schemas exposed in metadata for UI generation
- Removed deprecated Agency APIs (agency_chart, completions proxy, message_files)
- SendMessage no longer accepts my_primary_instructions field
- get_agency_structure deprecated, use get_agency_graph instead
- Conversation starters cache and metadata exposure
- Shared resource parameters for tools and MCP servers
- Local file path support in FastAPI with allowlisting