Instagram is Poor to agents.
Discry independently scored how well an AI agent can discover and understand the Instagram API from what’s public — not whether it’s usable. Below: every signal we checked, what’s costing the score, and what to change.
SCORED UNDER RUBRIC 1.2 · A full re-launch under Discry Score 2.5 — a new behavioral instrument, not comparable to these scores — is in progress.
Discovery
45% of score · 48/100Comprehension
55% of score · 53/100What we found
- Instagram API documentation lives within the broader Meta developer portal — an agent searching for 'Instagram API' must navigate through Meta's multi-platform documentation sprawl to find Instagram-specific endpoints
- Meta provides llms.txt at developers.facebook.com but it covers all Meta platforms (WhatsApp, Marketing API) with minimal Instagram-specific content, reducing its value for agents targeting the Instagram API specifically
- No official OpenAPI spec is published by Meta — third-party generated specs exist on Versori but are not maintained by the platform team, creating reliability concerns for agent tooling
- Community MCP servers exist across Smithery, PulseMCP, and Glama but none are official Meta-maintained servers, meaning agents depend on community maintenance for tool integration
- The robots.txt explicitly allows all crawlers with a broad 'User-agent: * Allow: /' directive, which is permissive for AI access
What to change
Prioritized by impact on discoverability. You (or your docs platform) deploy these — Discry never touches your API.
- 01Publish an official Instagram Graph API OpenAPI spec maintained alongside the quarterly API version releases
- 02Create an Instagram-specific llms.txt that covers the Instagram Platform APIs, scopes, and common workflows separate from the Meta-wide llms.txt
- 03Add llms-full.txt with comprehensive Instagram API documentation in markdown for LLM consumption
- 04Build or endorse an official Instagram MCP server to replace the fragmented community server ecosystem
- 05Add multi-step workflow guides showing common agent patterns: authenticate -> get media -> analyze insights -> moderate comments
Execution coverage · INFORMATIONAL, UNSCORED
Whether an agent can actually complete a call and recover from errors is the deeper Audit layer — documented here, but not part of the Discry Score.
OAuth2 with multiple scope types for different access levels. Graph API versioning (currently v21.0) with quarterly releases. Rate limits documented per-app and per-user. Cursor-based pagination. JSON error responses with error code, subcode, and message. No idempotency support documented.