Show off your agentic readiness — the badge auto-updates when your score changes.
llms.txt Content
# llms.txt for developers.facebook.com
# See https://llmstxt.org/ for specification
> Facebook Developer Documentation
This site contains developer documentation for Facebook platforms and APIs.
## Documentation Sections
- /docs/ - API and SDK documentation
- /blog/ - Developer blog and announcements
- /tools/ - Developer tools and utilities
- /products/ - Product documentation
## Additional Resources
- https://developers.meta.com - Meta Developer Platform
- https://ai.meta.com - Meta AI
- https://llama.meta.com - Llama models
## Meta Marketing API
Guidelines for AI agents and automated tools using the Meta Marketing API (Graph API) and Ads MCP Server.
### Agent Self-Identification
Include structured identification in your User-Agent HTTP header on all requests.
Format: <AgentName>/<Version> (<ModelName>) <HTTPClient>
Examples:
User-Agent: MyAdBot/1.3 (GPT-4o) python-requests/2.31.0
User-Agent: CampaignAgent/2.0 (Claude-Sonnet-4) axios/1.7.2
User-Agent: AdOptimizer/0.9 (Gemini-2.5-Flash) curl/8.4.0
Fields:
- AgentName (required): Your agent or product name. Be consistent across requests.
- Version (required): Semantic version of your agent.
- ModelName (recommended): LLM model powering the agent. Omit if unknown.
- HTTPClient (existing): Keep your HTTP library's default UA string.
Rules:
- Append agent info to the existing User-Agent. Do not replace it.
- Use the same agent name on every request. Do not randomize.
- Do not impersonate other agents.