docker

Chronos Docker

This directory hosts the Dockerfiles, compose files, and supporting code for running Chronos in containers.

Quickstart — demo stack (recommended)

The fastest way to evaluate Chronos. This compose will start Chronos and Postgres containers, then a simple seeder registers two MCP servers (Memory + Fetch) and visual agentflow with preset OpenRouter credentials

cd chronos/chronos_app/docker
docker build -f Dockerfile.demo -t chronos:demo ..

export OPENROUTER_API_KEY=sk-or-...
export OPENROUTER_LLM_MODEL=openai/gpt-4o-mini \

docker compose -f docker-compose.demo.yml up
# chronos is now accessible on http://localhost:3001
# login: admin@admin.com / test1234

After compose starts you get:

• An OpenRouter credential the canvas can attach to any Chat node.
• Two MCP servers registered as stdio (Memory via npx, Fetch via uvx). Both should reach HEALTHY within a few seconds.
• Logged seeder output confirms +N created / N skipped per category.

The seeder is idempotent: re-running docker compose -f docker-compose.demo.yml up will not duplicate records or overwrite manual edits made in the UI. The Dockerfile.demo image extends Dockerfile.local with uv so the Python-based fetch MCP preset works out of the box; production Chronos images stay lean.

Other deployment paths

Build and run Chronos locally, without Docker container images. Uses local Nodejs runtime enviroenment:

# nodejs runtime v24 is required. use nvm if necesary
# make sure you are in /chronos_app directory
git clone git@github.com:intelligexhq/chronos.git
cd chronos/chronos_app
node -v
pnpm install # pnpm nuke && pnpm install --frozen-lockfile
pnpm build # pnpm build --force
rm -rf ~/.chronos # chronos dev data stores. see https://intelligex.com/chronos/how-chronos-agent-builder-stores-data
# export ENABLE_AGENTS=true && export ENABLE_MCP_SERVERS=true
pnpm dev
# chronos is accessible on localhost:3000

Build and run a local Docker container image:

# clone & go to docker directory
git clone git@github.com:intelligexhq/chronos.git
cd chronos/chronos_app/docker
docker build -f Dockerfile.local -t chronos:local ..
docker run -d --name chronos -p 3001:3000 chronos:local
# docker run --name chronos2 -p 3003:3000 -e CHRONOS_INITIAL_USER=admin@admin.com:test1234:admin chronos:local
# docker run --name chronos2 -p 3003:3000 -e ENABLE_SCHEDULES=true -e CHRONOS_INITIAL_USER=admin@admin.com:test1234:admin chronos:local
# chronos is now accessable on http://localhost:3001

Use the local container image in docker compose:

docker compose -f docker-compose.yml up  # or docker compose -f docker-compose.yml up -d
# docker build -f Dockerfile.local -t chronos:local .. && docker compose -f docker-compose.yml up
# docker build -f Dockerfile.local -t chronos:local .. && docker compose -f docker-compose-vectordb.yml up
# docker build -f Dockerfile.local -t chronos:local .. && docker compose -f docker-compose-schedules.yml up
# docker build -f Dockerfile.local -t chronos:local .. && docker compose -f docker-compose-opentelemetry.yml up
docker compose ls
docker-compose down # or docker-compose down --volumes
# chronos is now accessable on http://localhost:3001

Worker mode with redis queues for horizontal scalability of agent request processing:

# run the docker compose which shows how to operate Chronos in queue / worker mode
docker compose -f docker-compose-workers.yml up
# scale workers if needed
docker compose -f docker-compose-workers.yml up --scale chronos-worker=3
# if enable you will see BullMQ dashboard at http://localhost:3001/admin/queues

Vector database mode for document embeddings example:

# run with Qdrant vector database and Ollama container for local embeddings
docker compose -f docker-compose-vectordb.yml up
# use ollama container and pull the embedding model after startup
docker compose -f docker-compose-vectordb.yml exec ollama ollama pull nomic-embed-text

# chronos is now accessible on http://localhost:3001
# configure vector store in UI: qdrant running at http://qdrant:6333
# configure embeddings in UI: ollama running at http://ollama:11434

Env Variables

To supply data to Chronos app during the startup you can use the following enviroenment variables - see .env.example

Compose files at a glance

File	Purpose
docker-compose.demo.yml	Recommended first run — seeded demo stack
docker-compose.yml	Basic single-service deployment
docker-compose.walkthrough.yml	Manual agent-registry walkthrough (see references/readme.md)
docker-compose.smoke.yml	Internal smoke test (CI / regression)
docker-compose-workers.yml	Horizontal-scale workers + Redis queues
docker-compose-vectordb.yml	Qdrant + Ollama for embeddings
docker-compose-schedules.yml	Scheduled-agent deployment
docker-compose-opentelemetry.yml	OTel + Jaeger observability stack

Need assistance?

We provide professional services to help you deploy, customise, and operate Chronos in your organisation’s environment:

• Architecture and deployment in your infra (on‑prem or cloud).
• Custom agent development, integrations of existing agents.
• MCP catalogues, discovery, auditing and security
• Training and best practices for teams.