Capsid
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Capsidsearch for documents about deployment"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Capsid
Capsid is a public, Cloudflare-native MCP server that serves a consolidated knowledge base from D1 and R2. It speaks MCP over Streamable HTTP and exposes a small, purposeful tool set: list, read, write, delete, move, find, search (FTS5), and namespaces.
Reads are open. Writes (write, delete, move) require an operator key sent as Authorization: Bearer <key>, verified against a sha256 hash stored as a Worker secret. Every write snapshots the prior version into document_versions and appends to audit_log, so you get history and rollback for free.
Stack
Cloudflare Worker (TypeScript), stateless MCP via
createMcpHandlerfrom the Agents SDKD1 for documents, versions, namespaces, and audit log, with FTS5 full text search
R2 (
MEDIAbinding) for mediaKV (
APP_KVbinding) for app state
Related MCP server: WikiJS MCP Server
Endpoints
POST /mcpMCP over Streamable HTTPGET /healthreturnsok, no auth
Clone setup
Install dependencies:
npm installCreate your own Cloudflare resources:
npx wrangler d1 create capsid npx wrangler kv namespace create APP_KV npx wrangler r2 bucket create capsid-mediaCopy the config template and fill in your IDs from step 2:
cp wrangler.jsonc.example wrangler.jsoncThe real
wrangler.jsoncis gitignored on purpose. Never commit it.Apply the migration:
npx wrangler d1 migrations apply capsid --remoteNote: the migration is idempotent (IF NOT EXISTS everywhere). Applying it against an already-migrated database, including the original capsid D1, is a no-op. Cloners must run it once.
Generate an operator key and store its hash as a secret. Keep the raw key somewhere safe; it is what MCP clients send as the bearer token.
npx wrangler secret put OPERATOR_KEY_HASHThe value must be the lowercase hex sha256 of your raw key. Never store the raw key anywhere in the repo.
Type check and deploy:
npx tsc --noEmit npx wrangler deployConnect an MCP client to
https://capsid.<your-subdomain>.workers.dev/mcp. Add theAuthorization: Bearer <key>header to enable writes.
Auth roadmap
v1 uses a single service token (bearer key checked against OPERATOR_KEY_HASH). The planned next step is human SSO via workers-oauth-provider, which slots in front of the same MCP handler. Not built yet.
License
MIT
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/DrDustinEdwards/capsid-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server