CAMEL-AI

CAMEL (camel-ai/camel on GitHub) is an open-source Python library for building multi-agent LLM systems: ChatAgents, role-playing societies, a 'Workforce' orchestration layer, memory/storage backends, RAG pipelines, and synthetic-data generation tooling (self-instruct, chain-of-thought distillation). It originated from the paper 'CAMEL: Communicative Agents for "Mind" Exploration of Large Language Model Society' (Li, Hammoud, Itani, Khizbullin, Ghanem; NeurIPS 2023), which is a genuinely early and heavily cited piece of multi-agent LLM research. The GitHub org describes CAMEL-AI as a community-driven research collective of 100+ researchers, not a company (https://arxiv.org/abs/2303.17760, https://github.com/camel-ai/camel).

17,646 stars, 2,059 forks, 483 open issues on the core repo as of 2026-08-27 (https://github.com/camel-ai/camel). It is actively developed: the most recent tagged release (v0.2.91a5) shipped 2026-07-13 with 40+ merged PRs from more than 20 first-time contributors, and the repo's last push was 2026-08-21 (https://api.github.com/repos/camel-ai/camel/releases, https://api.github.com/repos/camel-ai/camel). PyPI's 'camel-ai' package is on 0.2.90 with a broad set of optional extras (web, document, communication, data tools) rather than a single monolithic install (https://pypi.org/pypi/camel-ai/json). GitHub's security-advisories endpoint for the repo returns no published advisories, and the project has a working private vulnerability-disclosure process via GitHub Security (https://raw.githubusercontent.com/camel-ai/camel/master/SECURITY.md). No independent CVE or OSV entries were found for the 'camel-ai' PyPI package.

CAMEL ships an unusually large surface area for a single framework: dozens of built-in agent types (ChatAgent, CriticAgent, KnowledgeGraphAgent, RepoAgent, MCPAgent, and more), integrations for vector stores (Milvus, Qdrant, Weaviate, pgvector, FAISS), graph databases (Neo4j, Nebula), sandboxed code execution (Docker, E2B, Daytona), and connectors for Slack, Discord, Notion, Google Workspace, GitHub, and Stripe, all listed as installable extras in the PyPI metadata rather than bundled by default (https://pypi.org/pypi/camel-ai/json). It also underpins two other notable open repos from the same org: OWL, a workforce-automation framework with 20,104 stars, and Eigent, a desktop 'cowork' app with 15,128 stars (https://api.github.com/repos/camel-ai/owl, https://github.com/eigent-ai/eigent). That ecosystem — one research paper spawning multiple independently popular downstream projects — is a real signal of adoption within the agent-framework space.

CAMEL itself has no pricing — it's a free Apache-2.0 library. The company behind it, Eigent AI (legally 'EIGENT UK LTD' per its site footer), sells Eigent, a desktop multi-agent 'cowork' app built on CAMEL. Eigent's own code is open source (Apache 2.0, 15,128 GitHub stars) and free to self-host with your own API keys or local models. Eigent's hosted/cloud version has published pricing: Free (bring-your-own-key, 500 signup credits), Plus at $19.90/month (annual) or $24.99/month, and Pro at $99.99/month (annual) or $129.99/month, each with a 7-day trial; Team and Enterprise tiers are 'Coming soon' / contact-sales (https://www.eigent.ai/pricing, verified 2026-08-27). Eigent states 10% of every subscription funds CAMEL-AI's open-source research, an unverified vendor claim about fund allocation.

There is no dedicated 'About'/team page publicly listing CAMEL-AI's current staff, governance, or non-profit/corporate status beyond the paper's five academic authors (Guohao Li et al., largely affiliated with KAUST at publication time) and the GitHub org (https://arxiv.org/abs/2303.17760). No independent funding data for CAMEL-AI itself was found; Eigent AI's own funding (if any) is likewise not disclosed on its site or in any primary source located during this research — this should be read as 'not established,' not as evidence of no funding.