Open sourcePython · ≈61k stars · MIT
CrewAI
Role-based multi-agent teams with a simple mental model.
CrewAI popularized 'crews' — role-based multi-agent teams where each agent has a role, goal and backstory. Ships with an enterprise control plane for deploying, monitoring and governing agent crews.
Pros & cons
Each point cites its sources and says what kind of evidence backs it. A second reviewer tried to refute every claim before it was kept.
Pros
- User reportsUsers find it fast for prototyping, with role-based agents that are intuitive to set up.[1]
- User reportsSeparate agents, tasks and tools, plus YAML config for roles and models, keep behaviour changes out of code.[1]
- MetadataMIT licensed and released very often (1.15.33 on Oct 29, 2026).[1]
Cons
- User reportsUsers report the whole crew hanging when one agent or API call times out; one moved the pipeline to LangGraph.[1]
- User reportsThe early simplicity becomes a constraint: users report hard-to-reproduce drift on long tasks and a 'loose and underspecified' abstraction.[1][2]
- User reportsForum users report poor visibility into agent decisions, roughly 10-minute crew runs, and a virtualenv near 1 GB.[1]
- Best for
- Demos and internal multi-agent prototypes where speed to a first result matters more than determinism.
- Watch out for
- Timeouts, hangs and limited observability; plan your own monitoring before production.
Assessed Oct 29, 2026 · adversarially reviewedSuggest a correction
Activity
- Last push
- · yesterday
- Latest release
- 1.15.44
· 2d ago - Last 90 days
- 20+ releases
- · 2d ago
What's Changed Features Add capability to save files as versioned, shareable artifacts. Bug Fixes Fix inline agent and crew action references relative to the flow file. Stop duplicating kept separators when merging chunks. Isolate nested defaults from a shared flow declaration.…
- · 2d ago
What's Changed Bug Fixes Fix detection of multimodal support for Claude 5. Documentation Update snapshot and changelog for v1.15.42. Contributors @Vidit-Ostwal, @anubhavlal07, @joaomdmoura
- · 3d ago
What's Changed Features Make ChromaDB an optional extra Default Knowledge to local Qdrant and decouple storage from ChromaDB Pick tools for standalone agents and add them later Bug Fixes Store output_json values as JSON to prevent date issues in .json and output_file Give each…
- · 9d ago
What's Changed Features Allow approved private URL hosts. Bug Fixes Fix projects to depend on PyMuPDF 2.x. Drop a real storage URL from the URL read tests. Keep earlier step outputs across a human-feedback resume. Strip cache_breakpoint marker on the LiteLLM path. Documentation…
- · 10d ago
What's Changed Features Remove chromadb from the embedding function layer. Add exact historical state lookup by row id. Wire jsonl into rag data type discovery. Scaffold standalone agents as JSON by default. Add backend record lifecycle and pagination tests. Bug Fixes Run…
Checked against GitHub on Nov 21, 2026. Release notes are shortened to plain text.
The team behind it
Coral Gables, FL · Est. 2023
Multi-agent orchestration, from open-source framework to enterprise platform.
CrewAI popularized role-based multi-agent teams ('crews') with a simple mental model and ships an enterprise control plane for deploying, monitoring and governing them.
$18M Series A led by Craft Ventures (2024)
In research (5)
- Agent-Native by Design: A Programming Language for Agentic ApplicationsVerdict: Incremental
Nov 16, 2026 · Baseline framework; the Meeting benchmark is adapted from a CrewAI flow.
- EREBOTS 4.0: A framework for conversational assistants with A2A agents and MCP toolsVerdict: Incremental
Oct 26, 2026 · Cited as an alternative multi-agent framework.
Oct 26, 2026 · One of the seven frameworks surveyed for tool-output truncation behaviour.
- GEAR: Feature-Based Configuration and Code Generation for Agentic SystemsVerdict: Incremental
Oct 26, 2026 · Target framework with the highest feature coverage (31 of 32).
- Agentic Architecture of Information Systems. The AAM/5C Model: A Reference Model for Distributed AutonomyVerdict: Weak
Oct 21, 2026 · Included in the framework mapping table.
Neighbours on the workbench
TypeScript · ≈395k stars · MIT
openclaw/openclaw
The viral open-source personal AI assistant that really does things.
Maintained by OpenClaw Foundation
v2026.11.6releasedTypeScript · ≈285k stars · MIT
deepseek-ai/deepseek-harness
Everything is a Plugin — DeepSeek's open agent harness.
Maintained by DeepSeek
dsh-v0.2.2-rc.5 prereleasedTypeScript · ≈212k stars · Sustainable Use License (fair-code; GitHub reports NOASSERTION)
n8n-io/n8n
Fair-code workflow automation platform with native AI capabilities.
Maintained by n8n-io
n8n@1.123.100releasedConnections
Relations from the Harness Hub knowledge graph, as charted in the State of Harness reports.
