Documentation
Your classes, over MCP.
Syla is an MCP server over your own Canvas coursework: lectures, readings, assignments, announcements, deadlines. Connect it to ChatGPT, Claude, or any client that speaks MCP, and the assistant answers from your real course material, with the file and page it came from.
Canvas is read-only, always. No tool submits, posts, edits or deletes anything in Canvas. The one tool that writes at all,
save_academic_memory, saves a study note into Syla itself, never anywhere your school can see.The endpoint
https://usesyla.app/api/mcpStreamable HTTP, JSON-RPC 2.0, protocol 2025-06-18. Assistants connect with OAuth; your own code uses an agent token. One protocol implementation serves both, so anything ChatGPT can do here, your agent can do too.
Try it
Once connected, these are the shape of the product:
“What's due this week?”
“What did my professor say about the substitution effect? Cite the page.”
“What changed in my classes since yesterday?”
Where to next
Connect a client
Claude, ChatGPT, Claude Code, Cursor, VS Code, or any MCP client.
Auth, tokens & scopes
OAuth for assistants, agent tokens for your own code, and what each scope grants.
Tools
The 13 tools, generated from the definitions the server actually runs.
Prompt library
Copy-pasteable workflows: exam prep, weekly planning, catching up.
Security & privacy
Read-only Canvas, scoped access, and where the one write goes.
Troubleshooting
Connection failures, stale data, missing tools, lost tokens.
13 tools, one endpoint, no write access to Canvas. Machine-readable: /llms.txt and /llms-full.txt, and every page serves markdown at .md.