← All tools

Agent / Automation

CrewAI

Updates

  1. Feature3d ago· GitHub Releases

    1.15.44

    CrewAI 1.15.44 adds versioned, shareable file artifacts and documentation for using them in flows. The release also fixes path containment, vision detection, caching, chunk merging, structured outputs, and several agent and flow issues.

  2. Feature4d ago· GitHub Releases

    1.15.43

    CrewAI 1.15.43 fixes detection of multimodal support for Claude 5. It also updates the snapshot and changelog for version 1.15.42.

  3. Feature4d ago· GitHub Releases

    1.15.42

    CrewAI 1.15.42 makes ChromaDB optional, defaults Knowledge storage to local Qdrant, and allows tools to be added to standalone agents later. It also fixes execution limits, serialization, streaming, vector-store integrations, and UTF-8 handling.

  4. Feature10d ago· GitHub Releases

    1.15.41

    CrewAI 1.15.41 adds approved private URL hosts and preserves earlier step outputs when resuming after human feedback. It also updates the PyMuPDF dependency and fixes cache-breakpoint handling through LiteLLM.

  5. Feature11d ago· GitHub Releases

    1.15.40

    CrewAI 1.15.40 removes ChromaDB from the embedding layer, adds exact historical state lookup, JSONL discovery for RAG data, and JSON scaffolding for standalone agents. It also fixes async after_llm_call hooks and stops agents with unresolved CrewAI AMP apps.

  6. News14d ago· GitHub Releases

    1.15.39

    CrewAI 1.15.39 documents the apps_not_connected error returned by the /kickoff and /status endpoints. No code changes are listed.

  7. Feature14d ago· GitHub Releases

    1.15.38

    CrewAI 1.15.38 changes the default model to gpt-6-luna and adds spreadsheet URL processing, evaluation span tags, and a headless evaluation option. It also improves evaluation failure handling, API-key checks, environment-file permissions, and CLI exit behavior.

  8. Feature16d ago· GitHub Releases

    1.15.37

    CrewAI 1.15.37 adds spreadsheet-reading support for agents through crewai-files. It also supports scaffolding and loading standalone agent projects and updates RagTool documentation.

  9. Feature17d ago· GitHub Releases

    1.15.36

    CrewAI 1.15.36 adds experimental conversational agent sessions and prompts for missing API keys at startup. It also fixes async execution for synchronous tools, Responses API tool calls, sourcedAnswer handling, ChromaDB metadata, and project dependency configuration.

  10. News19d ago· GitHub Releases

    1.15.35

    CrewAI 1.15.35 updates the LiteLLM model catalog and adds release documentation for version 1.15.34. No code changes are listed.

  11. Feature21d ago· GitHub Releases

    1.15.34

    CrewAI 1.15.34 adds skill search, Python flow inputs, ZIP deployment creation, and run-stop tracking. It also improves provider and API-key errors, structured output handling, context-window behavior, repeated tool-call handling, Python 3.10 compatibility, and LiteLLM streaming errors.

  12. Feature25d ago· GitHub Releases

    1.15.33

    CrewAI 1.15.33 adds skill search, Python flow inputs, and ZIP-based deployment creation. It also fixes token accounting, compatibility, JSON crew loading, tracing configuration, authentication, checkpoint restoration, and flow behavior.

  13. News26d ago· GitHub Releases

    1.15.32

    CrewAI 1.15.32 fixes flaky tests and updates release documentation. No user-facing functionality is described.

  14. Feature26d ago· GitHub Releases

    1.15.31

    CrewAI 1.15.31 adds conversational flow context compaction and fixes multiple security issues, including deserialization, SQL injection, and path traversal vulnerabilities. It also improves guardrails, provider cleanup, storage handling, and model compatibility.

  15. FeatureOct 24· GitHub Releases

    1.15.30

    CrewAI 1.15.30 fixes AGENTS.md evaluation behavior so the login process starts correctly and stops after three rounds. It also updates the snapshot and changelog for version 1.15.29.

  16. FeatureOct 23· GitHub Releases

    1.15.29

    CrewAI 1.15.29 adds LiteLLM-compatible arguments, CLI application search, expanded tracing and evaluation tools, direct integration actions, and Azure DocumentDB vector search. AgentExecutor is now out of experimental status, alongside fixes for memory privacy, logging, replanning, and tool policies.

  17. FeatureOct 14· GitHub Releases

    1.15.28

    CrewAI 1.15.28 adds an evaluation command that identifies where stopped models failed and uses LiteLLM’s catalog for context-window data. The release also fixes human-feedback resumption, deployment diagnostics, streamed Anthropic tool events, RAG JSON decoding, and other platform issues.

  18. FeatureOct 9· GitHub Releases

    1.15.27

    CrewAI 1.15.27 adds XPU support for OpenCLIP, a DeepInfra OpenAI-compatible provider, and per-run cost and time tracking for model evaluations. It also improves evaluation records and fixes stop-sequence handling, source attribution, and SQLite initialization contention.

  19. FeatureOct 8· GitHub Releases

    1.15.26

    CrewAI 1.15.26 fixes interrupted read-write lock ownership, documentation source URL preservation, and task output logging so logs show the actual output.

  20. NewsOct 7· GitHub Releases

    1.15.25

    CrewAI 1.15.25 only adds documentation for the 1.15.24 snapshot and changelog. No functional changes are listed.

  21. FeatureOct 7· GitHub Releases

    1.15.24

    CrewAI 1.15.24 adds evaluation briefs, model and role swapping, Oracle integrations, queued background replies, and experimental job lifecycle features. It also improves evaluation exit behavior, human-feedback outputs, event reruns, dependency security, message summarization, and context-window handling.

  22. FeatureSep 28· GitHub Releases

    1.15.23

    CrewAI 1.15.23 adds native Gemini 3.8 Flash support, evaluation of the latest traced run, and richer tracing metadata. It also improves provider retry behavior, Bedrock async fallback, platform setup, and resource cleanup.

  23. FeatureSep 16· GitHub Releases

    1.15.22

    CrewAI 1.15.22 adds model routing by agent role, OpenRouter embeddings, richer tracing, platform integration validation, and application catalog access. It also fixes persistence, streaming tool calls, Gemini file content, Windows lock backups, and reasoning-effort handling.

  24. FeatureSep 9· GitHub Releases

    1.15.21

    CrewAI 1.15.21 adds checkpoint-runtime and CLI telemetry, alongside fixes for provider routing, streaming tool calls, JSON checkpoints, schema conversion, and integrations. It also corrects the documented GPT-4o mini context window to 128,000 tokens.

  25. FeatureSep 4· GitHub Releases

    1.15.20

    CrewAI 1.15.20 fixes discovery of legacy platform tool aliases and updates release documentation.

  26. IntegrationSep 4· GitHub Releases

    1.15.19

    CrewAI 1.15.19 adds a Clipper client, CEL now(), run-ending records, coarse machine-size reporting, and injectable platform-tool clients. It also fixes Gemini, Ollama, Claude, memory, URL-reading, and model-hook issues while updating vulnerable dependencies.

  27. FeatureAug 27· GitHub Releases

    1.15.18

    CrewAI 1.15.18 stabilizes conversational flows and expands declarative chat configuration, deployment tracking, and project telemetry. It also fixes tool-result preservation, conversation roles, task failure reporting, and Claude Sonnet 4.6 context and token defaults.

  28. FeatureAug 20· GitHub Releases

    1.15.17

    CrewAI 1.15.17 adds declarative conversational flows and improves message chunking. Fixes address OpenAI Responses API tool calls, MCP server naming, agent-scope cleanup, tool-error attribution, and SSRF checks across redirects.

  29. FeatureAug 14· GitHub Releases

    1.15.16

    CrewAI 1.15.16 adds UUID-based execution context management and expanded flow, trace, and deployment telemetry. It also fixes span release reporting, MySQL table-name validation, and failure-state leakage between turns, with new CopilotKit and AG-UI frontend guides.

  30. FeatureAug 12· GitHub Releases

    1.15.15

    CrewAI 1.15.15 adds reporting for flow outcomes, duration, and human-in-the-loop signals. It also fixes flow event and tracing behavior and updates Torch and GitPython for security.

  31. FeatureAug 11· Changelog

    CrewAI changelog updated

    CrewAI documentation now identifies the framework as version 1.15.14. The excerpt does not provide release notes or describe changes in this version.

  32. NewsAug 8· GitHub Releases

    1.15.14

    ## What's Changed ### Features - Split runtime context from coding agent and add project ID ### Documentation - Update snapshot and changelog for v1.15.13 ## Contributors @joaomdmoura

  33. NewsAug 7· GitHub Releases

    1.15.13

    ## What's Changed ### Bug Fixes - Fix preservation of provider on LiteLLM-routed models. - Harden brittle LLM event-bus mocks. - Fix underreporting of Anthropic cache token usage. - Bump h2 to version 4.4.1 to address security vulnerability GHSA-6hr6-w5qg-qmwg. ### Documentation - Add DOCS_TRANSLATIONS workflow for l

  34. NewsAug 5· GitHub Releases

    1.15.12

    ## What's Changed ### Features - Bump Flow canary on release - Add URLReadTool for reading arbitrary URLs - Add app metadata to platform action tools - Unify scaffolding under `crewai create <resource>` ### Bug Fixes - Clarify conversational route/handler name collision errors ### Documentation - Update scaffold AGE