A
licenseNot graded
qualityB
maintenanceEnables coding agents to offload quick judgment calls like shipping readiness, file triage, and claim verification to a fast local MCP server with calibrated confidence and safe fallbacks.
6
MIT
System-1 Gateway
by TheMattHub
This MCP server exposes three tools for fast System-1 decisions, health checks, and gateway introspection.
system1_decide: Ask the configured System-1 provider a question and returndecided,abstained, orerrorwith decision, confidence (0–1), rationale, and request_id. Supports optional context, 2–256 candidate options, caller-supplied request_id, min_confidence, and schema_version.system1_health: Check gateway and provider health asok,degraded, ordown.system1_capabilities: Describe the gateway, including version, schema versions, provider chain, thresholds, and JSON schemas.
Integrates Ollama as a decision engine provider, allowing the gateway to use local chat models for fast System-1 decisions alongside other providers.
Click on "Deploy 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., "@System-1 Gatewaywhich tool should I use to parse this JSON?"
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.
System-1 Gateway for Claude Code
The interactive version on the docs site shows each engine's numbers and each kind's own frontier on hover.
The System-1 Gateway gives Claude Code fast multiple-choice decisions from a small local engine, through one
MCP server with three stable tools: system1_decide, system1_health and system1_capabilities. You pick
the engine (Laya, Von, Liquid AI d1 through llama.cpp, any Ollama chat model, a Hugging Face reranker through a
small local server, a hosted Claude model, or one you add), and below its confidence threshold it abstains,
so Claude reasons the call through itself. Python 3.11+ and the standard library only; installed once at user
scope, it works in every project and with any MCP client.
Documentation: https://thematthub.github.io/Sys1-AI-Gateway/
Quick start
In Claude Code:
/plugin marketplace add TheMattHub/Sys1-AI-Gateway
/plugin install system1@system1-gateway
/system1:setup/system1:setup installs the gateway into ~/.system1-gateway/venv, runs the tests and registers the
system1 MCP server; /system1:choose-model recommends and installs an engine for your hardware. Restart
Claude Code, and Claude can call system1_decide. The
tutorial walks through it, and
Install the gateway covers the install
script and other MCP clients.
Related MCP server: laya-mcp
Which engine?
The naga decisions study compares every engine we tried on picking one of naga's 90 agent tools, on frozen blind test sets and one laptop GPU. Claude Haiku 5.5 is the most accurate on the search queries an agent writes (69% right first time, about 1 s), and a 150M gte-reranker trained on Claude Opus 5.5's labels comes within noise of it offline (67%) and beats it on people's own wording. Among published local models Gemma 4 E4B and Qwen3-Reranker-0.6B come next at 3 to 6 s, Laya is the fast pick when a wrong tool can be retried, and Liquid AI's d1-3B is the System-1 model level with the hosted ones on people's requests (13.7 s on 4 GB; commercial use below USD 10 M revenue). The study page has the numbers, the retries chart and the caveats.
Documentation
How-to guides: install, engines, configure, model routing, calibrate, add a provider, troubleshoot
Reference: tools and schema, config keys, slash commands and scripts, providers
Background: what System-1 means here, how routing and calibration work
Studies and release notes
For AI agents: llms.txt
Licence
Apache-2.0, see LICENSE and NOTICE. Contributions are welcome: see CONTRIBUTING.md; to report a vulnerability, see SECURITY.md.
Available Tools
3 toolssystem1_capabilitiesA
Describe the gateway: version, schema versions, provider chain, thresholds and JSON schemas.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
system1_decideA
Ask the configured System-1 provider for a fast decision. Returns status decided | abstained | error, with decision, confidence (0-1), rationale and request_id.
| Name | Required | Description | Default |
|---|---|---|---|
| context | No | Free-form structured context passed to the provider. | |
| options | No | Candidate answers. If given, the decision is exactly one of them. | |
| question | Yes | The decision to make, phrased as a question. | |
| request_id | No | Optional caller-supplied id, echoed back. Generated if omitted. | |
| min_confidence | No | Abstain below this confidence. Defaults to the configured threshold. | |
| schema_version | No | 1 |
system1_healthA
Check gateway and provider health (ok | degraded | down).
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
3 tool updates
v0.1.0- First observed
system1_capabilities - First observed
system1_decide - First observed
system1_health
TDQS
A3.9/5.0
Scored across 3 tools
Disambiguation5/5
Each tool targets a clearly distinct concern: making a decision, checking health, and introspecting capabilities. There is no overlap in purpose or resource, so an agent can select without confusion.
Naming Consistency5/5
All three tools share the system1_ prefix and snake_case verb/noun form (decide, health, capabilities). The pattern is predictable and uniform throughout.
Tool Count4/5
Three tools is on the lean side, but for a thin inference gateway (decide + health + capabilities) each earns its place. Slightly under-scoped if richer workflows are expected.
Completeness4/5
The core lifecycle of a decision gateway is covered: decision, health monitoring, and capability discovery. Minor gaps exist, such as batch decisions or recording decision outcomes, but agents can work around them.
Maintenance
ActivityNo data
ResponsivenessOutdated data
Related MCP Connectors
MCP server for your apps' tools and custom tools, plus hosted AI agents and approval-gated workflows
Decision Layer for AI Agents — 58+ tools, Advisor, MCP. Free key: POST /v1/register {}.
AI Reasoning Cache & Consensus Layer with 11 MCP tools via Streamable HTTP.
MCP server for progressive tool usage at any scale (see https://klavis.ai)
Related MCP Servers
- AlicenseAqualityBmaintenanceEnables MCP clients to run fast, local, non-autoregressive System-1 decisions on text, including classification, triage, routing, and safety gating, via six typed tools with probabilistic outputs.8Apache 2.0
- FlicenseNot gradedqualityAmaintenanceEnables Claude Code and other MCP clients to offload bounded work—code review, test stubs, boilerplate, commit messages, format conversion, embeddings—to a local GPU model or a cheaper cloud, OpenAI-compatible endpoint. Source files are read by the server rather than the main model, cutting token spend while the orchestrating model keeps planning and verifying.-
- AlicenseNot gradedqualityBmaintenanceEnables Claude Code and other MCP clients to offload bounded tasks such as code review, test stubs, commit messages, and format conversion to local or cheaper cloud LLMs, cutting token usage while the primary model remains the orchestrator.MIT