Skip to content

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Latest commit

 

History

14 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 

Repository files navigation

image

Couchbase Labs Co-Innovation

A curated index of Couchbase Labs projects built with and for customers — agentic applications, AI Data Plane tooling, migration / onboarding applications, and operational control planes, all running on Couchbase Server Enterprise Edition or Couchbase Capella.

Projects • Getting started • Contributing


What this repo is

Couchbase Labs co-innovation projects are reference implementations and demos that show how to build production-shaped AI and data applications on Couchbase. Each project lives in its own repository; this repo is the front door that explains what each one does, how they relate, and where to start.

Note: Couchbase Labs projects are experimental and community-supported. They are not part of the Couchbase product line and are not covered by Couchbase Support unless stated otherwise in the individual repository.

Projects

Project What it is Repository
Agent Operations Manager (AOM) Couchbase Agent Operations Manager (AOM) is a self-hosted control plane for AI agents, built on Couchbase Server Enterprise Edition that eases adoption of the Couchbase AI Data Plane. Migrating an agent's existing memory to Couchbase Agent Memory for Context Caching and LLM Caching is simplified, allowing existing and new agents to rapidly benefit couchbase-agent-operations-manager
Onboarding Agent Couchbase Onboarding Agent is a guided, AI-assisted bi-directional replication from MongoDB, DynamoDB, Redis, Cassandra, Cosmos DB, or another Couchbase cluster into Couchbase couchbase-onboarding-agent
Onboarding Agent Demo Databases Pre-seeded source databases for exercising the Onboarding Agent end to end couchbase-onboarding-agent-demo-databases
Optimizer Agent Couchbase Optimizer Agent is an AI agent with a dashboard UI that continuously analyzes Couchbase Enterprise and Couchbase Capella clusters, analyzes query and index optimization opportunities, and recommends optimizations. The agent runs real-time or from a support bundle, recommends optimized queries and index opportunities couchbase-optimizer-agent

Agent Operations Manager (AOM)

couchbase-agent-operations-manager

image

A self-hosted control plane that sits between AI agents and MCP tool servers. Agents authenticate to AOM, discover tools for a task, and invoke them; every invocation is re-authorized before it is proxied and every decision is written to an append-only audit log. AOM bundles the Couchbase AI Data Plane building blocks (MCP Server, Agent Memory, Agent Catalog) so client developers can build their own agents on top of it.

  • Tool governance — tool definitions embedded and stored in Couchbase; discovery combines RBAC, trust filtering, and vector search
  • Hijacking / drift detection — scans tool metadata at ingest, live responses on each invoke, and re-scans the catalog on a timer; flags poisoned-response → high-risk-tool chains
  • LLM gateway — one endpoint for Claude, ChatGPT, Gemini, and Databricks Model Serving with exact + semantic caching and savings dashboards
  • Memory & knowledge — per-user agent memory, context cache, and a role-filtered knowledge base for RAG
  • Oversight — agent identities with key rotation and OIDC, run traces, evaluations with a regression gate, human-approval tier, SIEM forwarding, and a web dashboard
  • Developer tools — Python SDK (aom_sdk), an MCP bridge, and integration skills for Claude, ChatGPT, and Gemini
  • Stack — Couchbase Server EE (vector search), FastAPI backend, React/TypeScript/Vite dashboard, Docker Compose, Helm chart for Kubernetes/K3s

Onboarding Agent

couchbase-onboarding-agent

image

Migrates MongoDB, Amazon DynamoDB, Redis, Apache Cassandra, Azure Cosmos DB, or another Couchbase cluster (CE, EE, or Capella) into Couchbase Server or Capella with a guided wizard, a live dashboard, and an AI assistant that runs on a local model.

  • Three migration modes for every source: one-time migration, continuous replication, and bulk copy + continuous sync with no gap between the two
  • Resumable checkpoints, precise rollback of the destination, and bi-directional replication (fallback sync after cutover, failback to resume forward sync)
  • Destination mapping of each source container to one or more bucket.scope.collection targets
  • Pre-flight validation — connectivity, versions, change-capture availability, keyspace collisions, TLS, delete propagation, rollback safety, XDCR vBucket compatibility
  • Live dashboard with throughput, replication lag, checkpoints, reconciliation results, and bottleneck findings, plus auto-throttling
  • Local assistant (Qwen 3 via an Ollama-compatible API) grounded in live migration state with a Couchbase-backed vector memory — credentials and data never leave your deployment
  • Stack — Docker Compose or Kubernetes (Helm); shares the AOM look and feel (light/dark mode)

Onboarding Agent Demo Databases

couchbase-onboarding-agent-demo-databases

Companion repository with pre-seeded source databases so you can run the Onboarding Agent end to end without pointing it at a production system.

Optimizer Agent

couchbase-optimizer-agent

image

An agent that inspects Couchbase workloads and recommends or applies optimizations — indexes, query shape, and cluster configuration — using the same AI Data Plane components the other projects build on.

Getting started

Every project ships with Docker Compose and, where applicable, a Helm chart. The general pattern is:

git clone https://github.com/couchbaselabs/<project>.git
cd <project>
cp .env.example .env        # add LLM API keys and Couchbase connection details
docker compose up --build

Each repository's README covers prerequisites, configuration, and production deployment. Start with the Agent Operations Manager if you want the control plane, or the Onboarding Agent if you want to move data into Couchbase first.

Common prerequisites

  • Docker and Docker Compose (or a Kubernetes cluster with Helm)
  • Couchbase Server Enterprise Edition 7.6+ or a Couchbase Capella cluster (several projects bundle a Couchbase container for evaluation)
  • An LLM API key (Anthropic, OpenAI, Google, or Databricks) or a local Ollama-compatible model, depending on the project

Contributing

Contributions, issues, and feature requests are welcome in the individual project repositories. To propose a new co-innovation project or to link an existing one from this index, open a pull request against this repo.

License

Each project is licensed independently — see the LICENSE file in its repository. The Couchbase Labs logo and the Couchbase name are trademarks of Couchbase, Inc.


Built by Couchbase Labs

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors