Jev directory
CommunityJev Agents & MCP
Jev MCP Server — glama.ai
wangkuangkuang's Python MCP bridge to Jev: three official question types plus compare, verify, and batch classify, with a one-command installer that writes client config for Claude Code, Pi, Cursor, OpenCode, or Codex.
- Reading time
- 2 min
- Published by
- Community
- Author
- wangkuangkuang
- Last verified
- Oct 2026
This is one of several Jev listings for the same project. The canonical entry is Jev MCP Server — mattioofr.
- Languages
- Python
- Capabilities
- Requires API key
Highlights
- Tools: choice (2-100 options), score (2-8 levels), noul, classify (up to 100 items), compare, verify, and a key setup command.
- Batch classification of 100 items costs about $0.007 and takes roughly a minute sequentially.
- Single calls return in about half a second and cost roughly $0.00002 to $0.0001 at $42 per 1B input tokens.
- Installs with uvx jev-mcp-server install claude-code (or pi, cursor, opencode, codex); a plain pip install works too.
- The setup tool verifies a TypeSafe key once and stores it with 0600 permissions; JEVMCP_CACHE=1 serves identical payloads from disk.
Watch out
The listing shows no licence, and it is a community project rather than an official TypeSafe server. It cannot explain its answers or handle open-ended reasoning, batch classification runs one call per item sequentially, and the latency and cost figures are the project's own indicative numbers.
Questions about this entry
Short answers drawn from the linked project or page.
An MCP bridge to Jev, TypeSafe's System One model, that lets clients make fast, cheap, calibrated structured decisions without an LLM. It needs a TypeSafe API key; there is no local model and nothing else to install.
Run uvx jev-mcp-server install claude-code (also pi, cursor, opencode, or codex) and the config is written for you; restart the client and ask a question. Without uv, install it via curl, brew, or pip, and the installer subcommand is the same either way.
compare runs an A/B choice with a visible probability split; verify returns the support degree of one claim against supplied evidence; classify batches up to 100 items against one shared category set with aggregated counts.
Examples measured by the project: a choice over 6 rich options plus context takes ~0.6-0.7s and ~1.7k input tokens for about $0.00007; a short noul or verify is ~$0.00002; classifying 100 items is about $0.007 at $42 per 1B input tokens.
Yes. Skip the env block and call the setup tool once from your agent: it verifies the key live and stores it locally with 0600 permissions. The env variable still beats the stored file when present.
With JEVMCP_CACHE=1 (or true/on), identical question payloads are answered from disk at zero cost with usage.cached: true, and classify deduplicates repeated items automatically. The cache is off by default, and its directory defaults to ~/.cache/jev-mcp.
Open-ended reasoning or long-context analysis, and anything that needs an explanation attached. Jev returns numbers; any reason an assistant adds is its own reading of those numbers, not output from the model.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: Jev in front of the HeyGen MCP
X: Playwright CLI + Jev vs Playwright MCP
X: A local MCP server that cuts computer use from 23s to 0.8s. Cartwright built Cortex, a local MCP server where…
X: A side panel that drives any site's WebMCP tools
More Jev resources like this
Long-term memory for AI agents whose reads and writes are driven by Jev's typed, probabilistic decisions instead of a generative LLM in the loop, inspired by the Jev-Mem paper and shipped as an MCP server, Claude Code hooks, a CLI and a Python library.
Jev Agents & MCP#community#github#memory
Community
illescasdaniel/jev-memPython
illescasdaniel/jev-memPythonA dependency-free MCP server that exposes TypeSafe's Jev to Claude Code, Codex, Cursor, Hermes, and other agents through the ask_jev and list_jev_models tools, so your code gets typed decisions with probabilities instead of prose to parse.
Jev Agents & MCP#community#catalog-import#logicrw
Community
codeia-academy/jev-mcpPython
codeia-academy/jev-mcpPythonUse Jev from Claude Code, Codex CLI, Claude Desktop or any MCP client, paid for by free Netlify credits and locked with a key only you have: a key-locked proxy forwards through Netlify AI Gateway and an MCP server exposes jev_evaluate and jev_models.
Jev Agents & MCP#community#catalog-import#logicrw
Community
demircigoksu/jev-netlify-mcpPython
demircigoksu/jev-netlify-mcpPythonAn agent skill for Claude Code, Codex and other coding agents, with a standard-library Python CLI, 5 tested question sets and findings from 19,367 Jev calls and 90.4M metered input tokens over 4 days.
Jev Agents & MCP#community#catalog-import#logicrw
Community
marcodicesare-dev/jev-skillPython
marcodicesare-dev/jev-skillPythonFreepik's dependency-light Go MCP server for Jev with five validated tools — decide, classify, verify, rerank, and list_models — OpenRouter or TypeSafe credentials, and Apache-2.0-licensed source.
Jev Agents & MCP#community#mcp#go
Community
A free Claude Code skill that puts Jev in front of the expensive model: thirteen commands sort, check, score, gate and verify with Jev, so Claude reads only the unsure slice and writes the output. Needs python3 and an OpenRouter key.
Jev Agents & MCP#community#python#claude-code
Community
charlesdove977/claude-x-jevPython
charlesdove977/claude-x-jevPythonFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.