Integrations
Connect Reza memory to coding agents through the CLI, hooks, local transcript adapters, and drop-zone ingestion.
Claude CodeHook
Native skill plus Stop hook flow
Cursor
Project transcript sync and rules
Codex
Local rollout sync and handoff commands
Aider
.aider.chat.history.md ingestion
Kilocode
Rules file and drop-zone fallback
GitHub Copilot
Copilot instructions plus context export
Continue.dev
Export or config-driven context
Codeium / Windsurf
Open context file in the editor
VS Code Extension
Search and graph UI inside VS Code
reza Integrations
reza works with every AI coding tool. Choose your tool below.
Supported Tools
| Tool | Integration Type | Guide |
|---|---|---|
| Claude Code | Native skill (auto-triggered) | claude-code/SKILL.md |
| Cursor | .cursorrules file | cursor/.cursorrules |
| Kilocode | Rules file | kilocode/rules.md |
| Aider | --read flag / .aider.conf.yml | aider/README.md |
| Continue.dev | @file / config.json | continue/README.md |
| GitHub Copilot | #file / copilot-instructions.md | github-copilot/README.md |
| Codeium / Windsurf | Context file | codeium/README.md |
| OpenAI Codex | System prompt / --read | codex/README.md |
Universal approach (works with any tool)
All tools can use reza via the exported context file:
# Generate the context file (run before each session):
reza export
# This creates .reza/CONTEXT.md — a compact markdown file
# that any LLM tool can read as context.Then give it to your tool:
- Aider:
aider --read .reza/CONTEXT.md - Cursor:
@.reza/CONTEXT.mdin chat - ChatGPT: paste the file content
- Any CLI tool:
cat .reza/CONTEXT.md | your-llm-tool
The universal workflow
Regardless of which tool you use, the workflow is the same:
# 0. One-time: install the Claude Code Stop hook (auto-syncs every turn, zero tokens):
reza install-claude-hook
# 1. Check for work from other AI tools:
reza session handoff --budget 8000 # structured brief ready to paste
# 2. Start your session:
reza session start --llm YOUR_TOOL --task "what you are doing"
# 3. Find files without scanning:
reza query --find "keyword"
# 4. Save progress:
reza session save --id TOOL-XXXXXXXX \
--summary "what was done" \
--context "key decisions, next steps, what failed"
# 4b. For tools other than Claude, ingest exported transcripts:
reza ingest .reza/handoffs/tool-20260410.json
# 4c. Recover older, relevant context on demand:
reza session search "keyword from the earlier discussion" --id TOOL-XXXXXXXX
# 5. End your session:
reza session end --id TOOL-XXXXXXXXAdding a new integration
PRs welcome! To add a new tool integration:
- Create
integrations/TOOLNAME/README.md - Cover: setup, how to inject context, cross-LLM handoff workflow
- Add the tool to the table in this README
See any existing integration as a template.