
Liquid
A universal adapter that lets Claude talk to APIs, databases, and industrial systems through a single interface. It auto-discovers schemas from REST endpoints, GraphQL, gRPC, databases (Postgres, MongoDB, Redis, Neo4j), MQTT brokers, Modbus PLCs, OPC UA nodes, and even other MCP servers. You get four operations: fetch, query, write, and sense (for live streams). The interesting bits are server-side filtering to avoid fetching 10,000 records when you need 10, cross-source normalization so Stripe and PayPal amounts look the same, and structured error recovery that tells Claude exactly what tool to call when auth breaks. Reach for this when you're tired of writing per-service connectors or when you need Claude to work with industrial hardware, IoT pub/sub, or a database you don't want to expose raw SQL to.
- tools
- 6
- auth
- Secret needed
- transport
- STDIO
- env vars
- 5
- last updated
- May 28, 2026
Tools
6 toolsPublic tool metadata for what this MCP can expose to an agent.
liquid_connectDiscover + map any interface.
Discover + map any interface.
No parameter schema in public metadata yet.
liquid_fetch
No parameter schema in public metadata yet.
liquid_queryServer-side search/aggregate.
Server-side search/aggregate.
No parameter schema in public metadata yet.
liquid_estimatePre-flight cost/size, no call.
Pre-flight cost/size, no call.
No parameter schema in public metadata yet.
liquid_list_adapters
No parameter schema in public metadata yet.
liquid_discover
No parameter schema in public metadata yet.
Sponsors
Configuration
OPENAI_API_KEYsecretOpenAI (or compatible) API key for discovery
OPENAI_BASE_URLOpenAI-compatible base URL, e.g. local Ollama/vLLM
GEMINI_API_KEYsecretGoogle Gemini API key (alternative)
ANTHROPIC_API_KEYsecretAnthropic API key (alternative)
LIQUID_ALLOW_WRITESSet to 1 to expose the liquid_execute write tool (database insert/update/delete). Off by default — read-only.
Readme
Liquid
Connect your AI agent to anything — with no connector to write or maintain.
Point Liquid at a URL or a database and it works out the interface for you:
discovers its shape, maps it to the fields you asked for, and handles auth,
pagination and normalization — typed records, no client code. When the upstream
drifts, it re-maps and keeps going. The same small API — fetch · query ·
write · sense — reaches web APIs, databases, other agents (MCP/A2A), email,
and even IoT and industrial systems (MQTT, Modbus, OPC UA, BACnet). An LLM does the
learning at setup (and on drift); the data path itself makes no model call.
What an agent can reach through Liquid
Details
- Repository
- github.com/ertad-family/liquid
- Category
- AI & LLM Tools
- Package
liquid-mcp- Transport
- STDIO
- Auth
- Required
- Registry
- MCP registry, active