# CrewAI 1.15.30: what's new, verdict & setup Canonical: https://theagenticweekly.com/tools/crewai · Released: 2026-10-23 · Last verified: 2026-10-25 · Evidence: Tested (smoke test, 2026-10-25, local Qwen2.5-3B-Instruct, not vendor models) · Vendor: CrewAI · Official site: https://github.com/crewAIInc/crewAI > Python framework for orchestrating role-playing, autonomous AI agents that work as a crew. ## Verdict **Reasonable if you like the crew-of-agents mental model; the memory and tool-loop fixes in this release are worth taking.** Good for: - Developers who want role-based multi-agent flows - Teams that want a built-in eval command - Users who need vector search on Azure DocumentDB Skip if: - You prefer an explicit graph model - You need a stable API with few releases; versions move quickly Tested. Tested 2026-10-25 (America/Chicago). "uv pip install crewai==1.15.30" in a fresh Python 3.13 venv worked; an earlier unpinned install into a shared venv resolved to 1.6.1, so pin the version. Ran one Agent, one Task and one Crew with a custom @tool and a local Ollama model (telemetry off): the agent used the tool and reported 9 words, which is correct. Not tested: the uv tool install / crewai CLI route, the eval command or Azure DocumentDB. Test environment: Linux x86_64 box with 8 CPUs and no GPU, no vendor account or API key. Agent tools were pointed at a local open-weight model (Qwen2.5-3B-Instruct, 4-bit, served by Ollama 0.40.11), so we tested the tool, not any vendor's hosted models. Extended 2026-10-25: CrewAI 1.15.30 passed our one-tool task on a local 3B model: 3 of 3 trials; cold uv install 8.4 s. Full log: /tests/crewai. Test scope: - Kind and date: smoke test, 2026-10-25 (America/Chicago). Not a benchmark. - What we ran: Installed crewai 1.15.30 in a fresh Python 3.13 venv and ran one Agent, one Task and one Crew with a custom tool; the agent used the tool and reported 9 words (correct). An earlier unpinned install resolved to 1.6.1. Extended: CrewAI 1.15.30 passed our one-tool task on a local 3B model: 3 of 3 trials; cold uv install 8.4 s. - Model used: Qwen2.5-3B-Instruct (4-bit), a small local open-weight model served by Ollama 0.40.11; no vendor-hosted model was used - Environment: Linux x86_64 box, 8 CPUs, no GPU; no vendor account, login or API key - Not tested: The uv tool install / crewai CLI route, the eval command, Azure DocumentDB, the hosted platform, hosted models; Any vendor-hosted model: everything ran on a local 3B model, so tool-call reliability with frontier models is not measured here; Multi-agent flows, memory, long conversations, streaming, tracing and cloud deployment; Anything beyond one tool and one question; three trials is a smoke test, not a benchmark. Full test log (commands, versions, raw output): [CrewAI 1.15.30: our test log](https://theagenticweekly.com/tests/crewai) Pricing: Not verified. We have not verified pricing for this tool; see the official site. ## Best for - [Best AI tool for agent building](https://theagenticweekly.com/best/agent-building) (#5): MIT framework with crew-style orchestration and a new eval command. ## What's new CrewAI 1.15.29 promotes AgentExecutor out of experimental status and adds LiteLLM-compatible kwargs, Azure DocumentDB vector search and new CLI evaluation features; 1.15.30 followed with a small fix on Oct 24. - 1.15.29 (Oct 23): AgentExecutor is no longer experimental. - 1.15.29: Support for LiteLLM-compatible kwargs; new Azure DocumentDB vector search integration. - 1.15.29: crewAI eval can run the crew itself for a verdict from one command and export evaluations as files. - 1.15.29: Fixes include private memories leaking into every recall, repeated identical tool calls, and LanceDB filter escaping. - 1.15.30 (Oct 24): Fix to the AGENTS.md login flow while evaluating. ## How to set it up: Install CrewAI with uv Install steps come from the vendor's docs. We ran the install and a short real task on 2026-10-25; see the test notes above for what was and was not covered. 1. **Install uv.** The README installs CrewAI with the uv tool installer, so install uv first. ``` curl -LsSf https://astral.sh/uv/install.sh | sh ``` 2. **Install CrewAI.** Install it as a uv tool and refresh your shell path. ``` uv tool install crewai uv tool update-shell ``` 3. **Check the install.** List your uv tools to confirm CrewAI is present, then follow the README's project scaffolding steps. ``` uv tool list ``` ## Pros and cons Pros: - MIT license and a large GitHub following - Built-in eval command - Fixes to memory and tool-loop behavior in this release Cons: - Very fast release cadence (several versions per week) - Some fixes address leaks and loops, which signals earlier rough edges - An unpinned install resolved to an older version (1.6.1) in our shared venv; we only ran a one-agent crew against a local model ## Alternatives - [LangGraph](https://theagenticweekly.com/tools/langgraph): Low-level framework for building long-running, stateful agents as graphs. - [OpenAI Agents SDK](https://theagenticweekly.com/tools/openai-agents-sdk): Lightweight Python framework for multi-agent workflows, with tools, handoffs, sessions and tracing. - [Mastra](https://theagenticweekly.com/tools/mastra): TypeScript framework for AI-powered applications and agents, with a local Studio. - [Langflow](https://theagenticweekly.com/tools/langflow): MIT-licensed visual builder for AI agents and MCP servers with a very large GitHub following. - [Botpress](https://theagenticweekly.com/tools/botpress): Customer-support AI agent platform with a free plan and conversation-based paid plans. - [Gumloop](https://theagenticweekly.com/tools/gumloop): Agent builder for teams; plans start at $37 per month with 20k included credits. - [Lindy](https://theagenticweekly.com/tools/lindy): AI 'employee' builder that connects to your tools; Free plan with $50 in credits that last 7 days, then credit-based plans. - [Microsoft Copilot Studio](https://theagenticweekly.com/tools/copilot-studio): Microsoft's platform for creating and managing custom agents, workflows and apps with no code. - [Pydantic AI](https://theagenticweekly.com/tools/pydantic-ai): Python agent framework built on Pydantic types, covering agents, realtime voice, image generation and embeddings. - [Relevance AI](https://theagenticweekly.com/tools/relevance-ai): Platform for specialist AI agents for sales, customer success, marketing and HR, with an embedded deployment team. - [Voiceflow](https://theagenticweekly.com/tools/voiceflow): Platform for chat and voice AI agents for customer experience teams. - [Vercel AI SDK](https://theagenticweekly.com/tools/vercel-ai-sdk): Vercel's TypeScript SDK that gives one API across many model providers, with streaming and agent helpers. ## FAQ ### What is new in CrewAI 1.15.29? AgentExecutor graduated from experimental, LiteLLM-compatible kwargs and Azure DocumentDB vector search were added, and crewAI eval gained one-command runs and exports. ### How do I install CrewAI? Install uv, then run uv tool install crewai and uv tool update-shell. ### Is CrewAI open source? Yes, the repository reports an MIT license. ## What we have not verified - Pricing of CrewAI's hosted platform (not covered) ## Sources - [CrewAI 1.15.29 release](https://github.com/crewAIInc/crewAI/releases/tag/1.15.29) (fetched 2026-10-25) - [CrewAI 1.15.30 release](https://github.com/crewAIInc/crewAI/releases/tag/1.15.30) (fetched 2026-10-25) - [README](https://github.com/crewAIInc/crewAI) (fetched 2026-10-25) ## How to cite Agentic Weekly, "CrewAI 1.15.30: what's new, verdict & setup", last updated 2026-10-25, https://theagenticweekly.com/tools/crewai Keep the Reported or Tested label when you quote the verdict.