Why route MCP tool selection through Jev?

With MCP, every tool from every connected server goes into the LLM prompt on every turn, which the README says reaches hundreds of tools and about 200k tokens before the request is read. Choosing tools is a decision, so a decision model can do it cheaper.

What does the jev-mcp-router benchmark show?

The repo's graphic states that without Jev the LLM reads 201,202 tokens of tool schemas per turn versus 981 with Jev, choosing tools is 32x cheaper and 37% faster than Claude Sonnet 5, and Jev picks every needed tool in 95% of requests versus 99%.

How much does Jev cost for routing?

The README says Jev takes a state plus typed questions and returns an answer per question, at about $0.04 per million input tokens, so routing hundreds of tools costs a fraction of a cent per request.

What scale does the jev-mcp-router example cover?

The README describes an example library of 649 tools, which is what makes the token comparison dramatic: the un-routed prompt carries every tool schema, while the routed prompt carries only the accepted handful.

How do I install jev-mcp-router?

Not documented in the digest excerpt, which shows the description, benchmark graphic and architecture diagram but no install or run commands; check the repository for current setup steps.

Who built jev-mcp-router and under what license?

The GitHub organization ini8labs publishes it under the MIT license. The repository was created 2026-10-04 and pushed the same day, with 4 stars and 1 fork when the sweep ran.

Jev directory

CommunityJev Integrations & Adapters

Jev MCP Router

An MCP gateway that routes tool selection through Jev: for each request, Jev answers one yes/no question per server and per tool, and the LLM only sees the tools that passed. The repo's benchmark covers 649 tools at 1/32 of a comparable routing cost.

See highlights Open github.com

Stars: 4
Language: Python
License: MIT
Reading time: 2 min
Published by: Community
Author: ini8labs
Last verified: Oct 2026

ini8labs/jev-mcp-router

Organization repository on GitHub

View on GitHub

Let a decision model pick your agent's MCP tools, not the LLM. An MCP gateway routed by Jev: 649 tools, 1/32 the cost of LLM routing.

Forks: 1
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.

Tagscommunity github mcp routing

Works forJev for Tool calling Jev for Routing Jev for Decision gating

Used inJev for Software & DevTools

Languages: Python
Platforms: MCP API
Capabilities: Open sourceRequires API key

Highlights

Jev answers one yes/no question per MCP server and per tool; the LLM sees only the tools Jev accepted.
The repo's benchmark: without Jev the LLM reads 201,202 tokens of tool schemas per turn; with Jev, 981.
Choosing tools claims 32x cheaper and 37% faster than Claude Sonnet 5, per the benchmark graphic.
It picks every needed tool in 95% of requests versus 99% for the LLM-routing comparison, per the same graphic.
Jev answers a state plus typed questions for about $0.04 per million input tokens, and the example library has 649 tools.

Watch out

README reviewed 2026-10-17; all numbers come from the repository's own benchmark image, the digest shows no install commands or independent testing, and pricing figures reflect Jev's listed per-token rate.

Questions about this entry

Short answers drawn from the linked project or page.

Reactions & coverage

Posts, threads, and videos about this entry from around the web.

Facebook: Intake Router keeps diagnosis out of the answer space at the front desk.

FacebookClinicnotes

X: OpenRouter Decision Model Rankings

X: herdr + jev, an agent router

X: A group chat bot with decision room routing

From the Jev guides

Original write-ups on what this entry is for.

Decision gates in agent frameworks: what each integration ships

Oct 16, 2026

How Pydantic AI, Mastra, TanStack AI, DSPy, Spring AI, and LiteLLM expose Jev: where the typed decision sits in the loop, what each integration returns, and which big frameworks are only adjacent.

Jev Integrations & Adapters#agents#frameworks#integrations

Read article

From the Jev AI model community

Posts from builders shipping with Jev right now.

Follow @typesafeai