
Voidfeed Mcp
Hooks Claude directly into VoidFeed's structured knowledge API without writing HTTP calls yourself. You get eight tools: catalog browsing, semantic search across content types, an agent directory with 523 entries including cost and failure modes, model benchmark lookups across 247 combinations, and fractal content retrieval with configurable depth. Free tier works immediately via npx with no auth. Paid tier unlocks full depth search, complete benchmark data, and a live community pulse tool. Payment happens through a Lightning invoice tool that returns a JWT, or via Stripe and USDC. Reach for this when you need Claude to compare models, discover agents, or pull structured knowledge graphs during reasoning instead of scraping docs.
- tools
- 7
- auth
- Secret needed
- transport
- STDIO
- env vars
- 2
- last updated
- May 22, 2026
Tools
7 toolsPublic tool metadata for what this MCP can expose to an agent.
voidfeed_catalogList all VoidFeed content types and discover what is available. Start here.
List all VoidFeed content types and discover what is available. Start here.
No parameter schema in public metadata yet.
voidfeed_get_contentRetrieve VoidFeed content by type and ID. Types: fractal, incomplete, signal, authority, capability.2 params
Retrieve VoidFeed content by type and ID. Types: fractal, incomplete, signal, authority, capability.
idstringtypestringfractal · incomplete · signal · authority · capabilityvoidfeed_semantic_searchSearch VoidFeed content semantically across all types.3 params
Search VoidFeed content semantically across all types.
limitintegerqstringtypesarrayvoidfeed_agent_directoryBrowse the AI agent directory with cost scores, orchestration roles, and failure modes.4 params
Browse the AI agent directory with cost scores, orchestration roles, and failure modes.
budget_constraint_usdnumbercapabilitystringlimitintegertask_typestringvoidfeed_benchmark_lookupLook up model benchmark scores across 247 model x benchmark combinations.2 params
Look up model benchmark scores across 247 model x benchmark combinations.
benchmarkstringmodelstringvoidfeed_model_compareCompare AI models on a task with cost, quality, and latency data.2 params
Compare AI models on a task with cost, quality, and latency data.
modelsarraytaskstringvoidfeed_enter_voidGet a Lightning invoice to unlock The Void tier (100k sats / 30 days full access).1 params
Get a Lightning invoice to unlock The Void tier (100k sats / 30 days full access).
content_typesarraySponsors
Configuration
VOIDFEED_JWTsecretJWT token for The Void tier (full depth access). Leave blank for free surface tier.
VOIDFEED_BASE_URLOverride the VoidFeed API base URL. Defaults to https://voidfeed.ai
Readme
VoidFeed MCP Server
Connect AI agents to VoidFeed — structured knowledge engineered for machine consumption.
Works with Claude Desktop, Cursor, Windsurf, and any Model Context Protocol compatible runtime.
What VoidFeed provides
Content
- Extract — structured datasets and depth-parameterized knowledge graphs (fractal)
- Recipe — verified technical how-to guides with load-bearing completion steps
- Fact-feed — citeable post-training-cutoff facts (model releases, protocol changes, API deprecations)
- Bazaar-intel — live reliability and routing data for x402 Bazaar endpoints
On-demand tools (per-call pricing, no subscription needed)
- Compose — grounded answers synthesized from internal knowledge stores via gpt-4.1 ($0.25)
- Market price — real-time crypto prices via CoinGecko/CoinCap ($0.10)
- Market safety — smart contract risk analysis via GoPlus ($0.10)
- Bazaar route — ranked endpoint recommendations for agent routing ($0.15)
Surface tier is free (no auth required). Premium tier unlocks full depth — pay via USDC on Base (x402, single request) or Stripe.
Details
- Repository
- github.com/voidfeedai-ops/voidfeed-mcp
- Categories
- AI & LLM Tools, Documents & Knowledge, Search & Web Crawling
- Package
voidfeed-mcp- Transport
- STDIO
- Auth
- Required
- Registry
- MCP registry, active