{"data":{"slug":"camel-ai-owl","name":"owl","tagline":"Optimized Workforce Learning for General Multi-Agent Assistance","github_url":"https://github.com/camel-ai/owl","owner":"camel-ai","repo":"owl","owner_avatar_url":"https://avatars.githubusercontent.com/u/134388954?v=4","primary_language":"Python","stars":20085,"forks":2299,"topics":["agent","artificial-intelligence","multi-agent-systems","task-automation","web-interaction"],"archived":false,"github_pushed_at":"2026-08-14T17:34:50+00:00","maintenance_label":"Very active","stars_delta_30d":116,"url":"https://www.graphcanon.com/tools/camel-ai-owl","markdown_url":"https://www.graphcanon.com/tools/camel-ai-owl.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/camel-ai-owl","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=camel-ai-owl","description":"🦉 OWL: Optimized Workforce Learning for General Multi-Agent Assistance in Real-World Task Automation","homepage_url":null,"license":null,"open_issues":116,"watchers":129,"ai_summary":"OWL is a framework designed to facilitate task automation through generalized multi-agent assistance, focusing on optimizing workforce learning and real-world interaction. It leverages the CAMEL framework for enhanced stability and performance.","readme_excerpt":"### Install Python\n\nBefore installing OWL, ensure you have Python installed (version 3.10, 3.11, or 3.12 is supported):\n\n> **Note for GAIA Benchmark Users**: When running the GAIA benchmark evaluation, please use the `gaia58.18` branch which includes a customized version of the CAMEL framework in the `owl/camel` directory. This version contains enhanced toolkits with improved stability specifically optimized for the GAIA benchmark compared to the standard CAMEL installation.\n\n```bash\n\n---\n\n# If not installed, download and install from https://www.python.org/downloads/\n\n---\n\n## **Installation Options**\n\nOWL supports multiple installation methods to fit your workflow preferences.\n\n---\n\n# Install uv if you don't have it already\npip install uv\n\n---\n\n# Create a virtual environment and install dependencies\nuv venv .venv --python=3.10\n\n---\n\n# Install CAMEL with all dependencies\nuv pip install -e .\n```\n\n---\n\n# Install from requirements.txt\npip install -r requirements.txt --use-pep517\n```\n\n---\n\n# Option 1: Install as a package (recommended)\npip install -e .\n\n---\n\n# Option 2: Install from requirements.txt\npip install -r requirements.txt --use-pep517\n```\n\n---\n\n### Option 4: Using Docker\n\n#### **Using Pre-built Image (Recommended)**\n\n```bash\n\n---\n\n# This option downloads a ready-to-use image from Docker Hub\n\n---\n\n# For users who need to customize the Docker image or cannot access Docker Hub:\n\n---\n\n# Make the script executable and build the Docker image\nchmod +x build_docker.sh\n./build_docker.sh\n\n---\n\n### Model Requirements\n\n- **Tool Calling**: OWL requires models with robust tool calling capabilities to interact with various toolkits. Models must be able to understand tool descriptions, generate appropriate tool calls, and process tool outputs.\n\n- **Multimodal Understanding**: For tasks involving web interaction, image analysis, or video processing, models with multimodal capabilities are required to interpret visual content and context.\n\n#### Supported Models\n\nFor information on configuring AI models, please refer to our [CAMEL models documentation](https://docs.camel-ai.org/key_modules/models.html#supported-model-platforms-in-camel).\n\n> **Note**: For optimal performance, we strongly recommend using OpenAI models (GPT-4 or later versions). Our experiments show that other models may result in significantly lower performance on complex tasks and benchmarks, especially those requiring advanced multi-modal understanding and tool use.\n\nOWL supports various LLM backends, though capabilities may vary depending on the model's tool calling and multimodal abilities. You can use the following scripts to run with different models:\n\n```bash\n\n---\n\n### **Install Playwright MCP Service**\n\n```bash\nnpm install -g @executeautomation/playwright-mcp-server\nnpx playwright install-deps\n```\n\nTry our comprehensive MCP examples:\n\n- `examples/run_mcp.py` - Basic MCP functionality demonstration (local call, requires dependencies)\n- `examples/run_mcp_sse.py` - Example using the SSE protocol (Use remote services, no dependencies)\n\n---\n\n# 📄 License\n\nThe source code is licensed under Apache 2.0.","github_created_at":"2025-03-03T15:42:51+00:00","created_at":"2026-07-07T17:38:36.680155+00:00","updated_at":"2026-08-19T06:00:47.744653+00:00","categories":[{"slug":"ai-agents","name":"AI Agents","url":"https://www.graphcanon.com/categories/ai-agents","markdown_url":"https://www.graphcanon.com/categories/ai-agents.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/ai-agents"},{"slug":"evaluation-observability","name":"Evaluation & Observability","url":"https://www.graphcanon.com/categories/evaluation-observability","markdown_url":"https://www.graphcanon.com/categories/evaluation-observability.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/evaluation-observability"}],"tags":[{"slug":"agent","name":"agent"},{"slug":"multi-agent-systems","name":"multi-agent-systems"},{"slug":"task-automation","name":"task-automation"},{"slug":"web-interaction","name":"web-interaction"}],"trust":{"provenance":{"is_fork":false,"github_id":942114501,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-08-19T06:00:46.954Z","maintenance":{"label":"Very active","score":96,"methodology":"github_public_v1","releases_90d":0,"days_since_push":4,"last_release_at":null,"stars_delta_30d":116,"open_issues_delta_30d":1},"security_summary":{"status":"findings","scanner":"osv@v1","low_count":96,"high_count":0,"last_scan_at":"2026-07-11T11:12:43.005Z","medium_count":0,"scan_profile":"deps","critical_count":0}},"capability_facts":{"scan":{"source":"repo_scan","observed_at":"2026-08-19T06:00:47.451Z"},"languages":{"value":["python"],"source":"github.language+pyproject.toml","observed_at":"2026-08-19T06:00:47.451Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":{"notes":["For customizing the Docker image or accessing through Docker Hub, ensure that scripts are made executable and use `chmod +x build_docker.sh` followed by `./buil","d_docker.sh` to build."],"requires_docker":true},"constraints":{"requires_docker":true},"when_to_use":["When you specifically require enhancements from the customized CAMEL framework version as provided in the `gaia58.18` branch for GAIA benchmark evaluation.","For situations where advanced multi-modal understanding and tool calling capabilities are critical, especially when using OpenAI models (GPT-4 or later versions) for optimal performance."],"when_not_to_use":["If you need a solution that operates outside of Python 3.10, 3.11, or 3.12 environments.","When the project does not require real-world task automation with multi-agent systems and focuses on simpler tasks without complex tool interactions."],"source":"enrich:decision_facts","observed_at":"2026-07-11T14:17:48.348Z"},"constraint_facets":{"requires_docker":true},"decision_summary":[{"label":"Requirements","value":"Requires Docker; For customizing the Docker image or accessing through Docker Hub, ensure that scripts are made executable and use `chmod +x build_docker.sh` followed by `./buil; d_docker.sh` to build."},{"label":"Adopt for","value":"- **when_to_use**: Ideal for scenarios where you need robust task automation through multi-agent systems with a focus on workforce learning and real-world interaction, especially if performing complex tasks or benchmarks"},{"label":"License detail","value":"The source code is available under an Apache 2.0 license."}]}}