Skip to main content
Glama

MCP Bluesky Stock Sentiment

architecture.md1.53 kB
# Architecture Overview This diagram shows how the Bluesky Stock Sentiment MCP server is wired end‑to‑end. ```mermaid flowchart TD A["MCP Client\n(e.g., Claude Desktop)"] -->|"HTTP MCP (/mcp)\nStreamableHttpTransport"| B["Cloudflare Worker\n(Hono + mcp-lite)"] B -->|invoke tool| C["Tool: analyze_stock_posts"] C -->|search terms| D["Bluesky Search API\n(@atproto/api)"] C -->|filter text| E["Stock filter\n(keywords/regex)"] C -->|analyze text| F["Workers AI\n(@cf/huggingface/distilbert-sst-2-int8)"] F -->|label + score| C C -->|summary text| B B -->|MCP content| A %% Optional real-time collection (reference implementation) C -. optional .-> G[("Bluesky Firehose\n(WebSocket)")] classDef node fill:#eef,stroke:#58a,stroke-width:1px,color:#111; classDef service fill:#efe,stroke:#5a8,stroke-width:1px,color:#111; class A,B,C,E node; class D,F,G service; ``` Key files - `src/index.ts` — Hono app, MCP server wiring, and routes. - `src/mcp/tools.ts` — Registers `analyze_stock_posts` and implements handler. - `src/lib/bluesky-api.ts` — Queries Bluesky search for recent posts. - `src/lib/stock-filter.ts` — Heuristic keyword/regex filtering for stock posts. - `src/lib/sentiment.ts` — Workers AI sentiment classification. Notes - The MCP endpoint is served at `/mcp` using `mcp-lite` over HTTP. - Workers AI runs the `@cf/huggingface/distilbert-sst-2-int8` text classification model by default. - The firehose collector is an optional reference and not wired into the Worker route.

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/Tar-ive/mcp-bluesky-stock-sentiment'

If you have feedback or need assistance with the MCP directory API, please join our Discord server