# Kindex

Kindex gives AI agents a shared project knowledge graph. Use it to carry decisions across sessions and coordinate agents working on the same project.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 47
- Author: wandercom
- License: MIT · policy=link-only · description only; the file stays with the author
- Source: https://github.com/wandercom/kindex
- Updated: 2026-10-23
- HTML: https://affpapa.org/en/nevibecode/mcp/kindex

## What it does

Kindex links project knowledge in a graph that agents can read and update. It keeps decisions and constraints available across sessions, and lets agents pass work between tools. Project data lives in the git-tracked `.kin/` directory. Several agents can use the shared context at the same time.

Install the Python package with the `mcp` extra, then create a graph with `kin init`. Optional components add Anthropic-based extraction through `llm` and semantic search through `vectors`. Python 3.10 or newer is required for the `uvx` setup. Without the `mcp` extra, the `kin` command alone does not provide the `kin-mcp` server.

## Requirements
- Needs: Python and uv (commands run via uvx)

## Install

### Claude Code (native platform)

**Platform CLI**
```bash
claude mcp add kindex -- uvx kindex
```
uvx runs the command named after the package; if the README names another one, use uvx --from kindex <command>

### Codex

**Platform CLI**
```bash
codex mcp add kindex -- uvx kindex
```
uvx runs the command named after the package; if the README names another one, use uvx --from kindex <command>

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "kindex": {
            "command": "uvx",
            "args": [
                "kindex"
            ]
        }
    }
}
```
uvx runs the command named after the package; if the README names another one, use uvx --from kindex <command>

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "kindex": {
            "type": "stdio",
            "command": "uvx",
            "args": [
                "kindex"
            ]
        }
    }
}
```
uvx runs the command named after the package; if the README names another one, use uvx --from kindex <command>

### Gemini CLI

**Platform CLI**
```bash
gemini mcp add kindex uvx kindex
```
uvx runs the command named after the package; if the README names another one, use uvx --from kindex <command>

## When to use it
- carry project decisions between sessions
- share context when handing work to another agent
- let several agents access the same project knowledge
- select session context against an available token budget

## Example prompts
- Save this decision and its related constraint in the project knowledge graph.
- Pass the current task context and decisions to the next agent.
- Find project knowledge relevant to this session within the available token budget.

## Similar
- [understand-onboard](https://affpapa.org/en/nevibecode/skills/understand-onboard.md) — skill · 3700 installs
- [understand-diff](https://affpapa.org/en/nevibecode/skills/understand-diff.md) — skill · 3600 installs
- [build-graph](https://affpapa.org/en/nevibecode/skills/build-graph.md) — skill · 1300 installs
- [gitnexus-cli](https://affpapa.org/en/nevibecode/skills/gitnexus-cli.md) — skill · 870 installs
- [dossier-collect](https://affpapa.org/en/nevibecode/skills/dossier-collect.md) — skill · 701 installs
- [understand-quickly](https://affpapa.org/en/nevibecode/mcp/understand-quickly.md) — MCP server · ★ 48

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md