MCP Server · Research CLI · Data Warehouse
Social Listening MCP Live
Ask what X, Reddit, TikTok and seven more platforms are saying about anything — through the sv research CLI, or the same 26 citation-returning tools plugged straight into Claude Desktop and claude.ai Research over MCP. OAuth 2.1, dual transport, Gemini video understanding, and an idempotent pgvector warehouse with write-through durability.
FastMCPPostgrespgvectorGeminiPython
Specifications
cli sv · 26 research tools
clients Claude Desktop · Code · Research
transport stdio + streamable HTTP
auth OAuth 2.1 · Google proxy
platforms 10 · normalized collectors
warehouse Postgres + pgvector · snapshots
video Gemini multimodal understanding
endpoint mcp.apekode.com
$ sv search_social query="tomo ai app"
[social-mcp] search_social(query='tomo ai app', channels=['x','reddit','github','youtube','tiktok','pinterest'])
[social-mcp] search_social -> returned 6/120 items in 10.9s errors={}
{ "query": "tomo ai app", "count": 6, "fetched": 120, "items": [
{ "channel": "tiktok", "engagement": 6752, "bookmarks": 2332, "views": 85621, ... },
{ "channel": "x", "author": "10xmylife", "views": 273954, "replies": 266, ... },
... ] }
sv CLI all speak to the same warehouse.
