Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up

kody

Multi-user MCP 'agents' cloud' on Cloudflare Workers with an optional search stage that reranks candidates with Workers AI Jev scores.

MCP serverbetaCustom license
View on GitHub ↗

Overview

Kody is described as your agents' cloud: the memory, keys, code and automations an AI agent keeps, portable across every MCP host. It is a multi-user personal assistant built on Cloudflare Workers and the Model Context Protocol, where each user gets an isolated assistant with packages, jobs, secrets and memories. It favours a compact MCP surface with powerful search and Code Mode execute flows over a large static tool catalog, and ships a Remix UI, OAuth-protected MCP endpoints and Worker-based routing in an Nx monorepo.

Its Jev use, per the listing, is an optional second search stage: widen the hybrid search pool, then rerank candidates with Score questions using typesafe/jev on Workers AI. The README excerpt focuses on the platform rather than this stage, so details of the reranking are not confirmed here. It is maintained by Kent C. Dodds under a custom licence.

What the decision model decides

TypeQuestionAnswers
ScoreHow relevant is this candidate to the search query?—

How it works

  1. Agent calls Kody's search over MCP
  2. Hybrid search returns a widened candidate pool
  3. Optionally, Jev on Workers AI scores each candidate
  4. Candidates are reranked and returned

Strengths

  • Optional, so search works without it
  • Runs on Cloudflare Workers AI
  • Portable across MCP hosts

Caveats

  • Reranking stage not detailed in the README excerpt
  • Custom licence
  • Large platform; Jev is a small part

Getting started

npm install && npm run dev (dev server at localhost:3742).

At a glance

Type
MCP server
Status
beta
Language
TypeScript
Runtime
Cloudflare Workers, MCP server
Has code
yes
Has tests
yes
License
Custom license

Use cases and integrations

How curated lists describe it

Frequently asked questions

How does Kody use Jev?

As an optional second search stage that reranks candidates with Score questions via Workers AI.

What is Kody?

A multi-user MCP platform holding an agent's memory, keys, code and automations.

Is the Jev stage required?

No, it is optional per the listing.

Related projects

  • hippo-memory

    Local long-term memory for coding agents with an optional Jev reranker that orders recalled memories.

    ProjectproductionTypeScript
  • dgui-hypermem

    Self-hosted memory MCP server on Cloudflare Workers where a Jev layer judges memory type, salience and durability and re-ranks recall.

    MCP serverprototypeTypeScript
  • apple-rag-mcp

    MCP server over Apple developer docs and WWDC transcripts that uses hybrid search, then Jev relevance scores to rank results.

    MCP serverproductionTypeScript
  • OpenViking

    Volcengine's context database for agents, with a Jev rerank client that scores candidate documents with jev-latest as relevance.

    LibraryproductionPython
  • BGTS Context Engine

    Deterministic code-graph context for coding agents on Postgres, AGE and pgvector, with optional Jev or local re-selection of symbols.

    MCP serverbetaPython
  • jev-rerank

    Agent skill and Python library for Claude Code and Codex that reranks candidate lists with one Jev yes/no relevance question each.

    ProjectbetaPython