freshcrate
Home > MCP Servers > chronos

chronos

Chronos is visual AI agent builder - tailored for self-hosted deployments and observability

Description

Chronos is visual AI agent builder - tailored for self-hosted deployments and observability

README

Chronos – visual AI agent builder for self‑hosting

Build Status License Node Version PRs Welcome LangChain Core LangGraph MCP SDK

Build, run, and debug AI agents on your own infrastructure — with full visibility and control. The self-hosted, production-ready alternative to SaaS AI agent builders

 

Chronos is a visual AI agent workflow builder designed for teams who own their infrastructure.

  • Visual builder for complex agent workflows
  • Fully self-hosted (local, on-prem, private cloud)
  • Deep observability (trace every step of every run)

TL;DR: Chronos is like Flowise — but tailored for teams who need self-hosting, observability, and production readiness.

Why Chronos?

Most AI agent tools are great for demos — but fall apart in production. Chronos is built and used for real-world applied AI use cases

  • Own your data: Run everything inside your infrastructure.
  • Visibility of agents actions: Inspect prompts, tool calls, responses, errors, and token usage.
  • Cost & Performance Dashboard: Track and optimise agent costs and operations.
  • Scheduled agents: Maintain schedules for your agents.
  • Tracing: Opentelemetry helps traicing failures across the entire workflows.
  • From prototype to production: Build agents and integrate to enterprise systems via embeded OpenAI APIs.

What can you build?

  • Internal copilots over private company data
  • RAG pipelines with local embeddings and vector stores
  • Automated reporting and monitoring agents
  • Deep research and analysis workflows
  • Scheduled background agents for ongoing work.

Quick start

Get Chronos running locally in under 5 minutes. The fastest way to try it is to build and run all‑in‑one Chronos Docker image.

# clone the repository
git clone git@github.com:intelligexhq/chronos.git
cd chronos_app/docker

# build local image
docker build -f Dockerfile.local -t chronos:local ..

# run Chronos
docker run -d --name chronos -p 3001:3000 chronos:local
# chronos is now available at:
# http://localhost:3001

For more configuration and advanced hosting examples, including localy hosted vectorstores, Ollama, queue modes - see Docker compose examples section.

Get involved

  • ⭐ Star this repo to support development
  • Contribute via PRs; sugest issues, features and use cases.

Backlog

Some of the upcoming capabilities:

  • Agents with webhooks for event‑driven workflows.
  • Versioning and publishing for agents.
  • SSO (single sign‑on) for user logins.
  • Additional tutorials on intelligex.com covering skill catalogues, deep agents, and data privacy.

Tutorials

These guides will get you started with running Chronos and building agents:

License

Source code in this repository is made available under the Apache License Version 2.0.

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 and integrations.
  • Training and best practices for teams.

Release History

VersionChangesUrgencyDate
v1.4.0- new langraph deep agent nodes available in visual builder. - scheduler engine for automating agent runs. allows to kick off agents on the regular basis. - cost & performance dashbord. analytics view for the token consumption & cost per agent / all running agents. - webhook support for agent tasks.Medium3/22/2026

Dependencies & License Audit

Loading dependencies...

Similar Packages

tenuoHigh-performance capability authorization engine for AI agents. Cryptographically attenuated warrants, task-scoped authority, verifiable offline. Rust core.v0.1.0-beta.22
obsidian-vault-intelligenceObsidian vault intelligence9.2.8
claw-pilotMulti-agent orchestration runtime with task board, flow engine, budget control, MCP integration and real-time dashboard. Self-hosted on Linux/macOS.v0.81.1
trace-mcpMCP server for Claude Code and Codex. One tool call replaces ~42 minutes of agent explorationv1.28.0
@contractspec/lib.ai-agentAI agent orchestration with MCP and tool support8.0.10