Telegram is Poor to agents.
Discry independently scored how well an AI agent can discover and understand the Telegram 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 · 19/100Comprehension
55% of score · 71/100What we found
- Telegram Bot API documentation is self-contained and well-organized with task-oriented descriptions and realistic examples, despite having zero AI-specific discovery signals
- The llms.txt URL returns a "Page not found" HTML page (200 status but no valid content) — deceptive for automated discovery tools
- Multiple community MCP servers exist across PulseMCP, Glama, and Smithery for bot messaging, channel access, and chat management
- No OpenAPI spec exists despite the Bot API being well-structured enough to generate one
- Documentation is HTML-only with no markdown alternative, limiting agent consumption quality
What to change
Prioritized by impact on discoverability. You (or your docs platform) deploy these — Discry never touches your API.
- 01Fix /llms.txt to either serve valid content or return a proper 404 status code
- 02Generate and publish an OpenAPI spec from the Bot API documentation
- 03Add a sitemap.xml covering Bot API and platform documentation pages
- 04Add llms.txt covering Bot API capabilities, methods, and webhooks
- 05Create markdown versions of API documentation for better LLM consumption
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.
Telegram Bot API uses bot tokens from BotFather. JSON error responses with error_code and description. Rate limits documented (30 msg/sec to groups). Offset-based pagination for getUpdates.