longsurf-ai / openchart
git clone https://github.com/longsurf-ai/openchart.git
v0.1.20 · 3d ago
OSS trading workspace with your favorite agents
—
stars, 24h
—
stars, 7 days
+259
stars, 30 days · #1,000
259
★ total
27
forks · 5 open issues
Star history starts 2026-10-20; the chart appears after the second daily snapshot.
LanguageTypeScript
License—
Created2026-10-05 (15d ago)
Last push2d ago
Watchers2
OwnerOrganization
GitHub topics
agentagentic-traderaichartingchartscharts-jsharnessllmllm-agentspinescriptpinescript-v6quantitativequantitative-financetradertrading-toolkits
Links
- GitHub: github.com/longsurf-ai/openchart
- Homepage: https://chart.chat
- Releases · Issues · Hacker News mentions
README excerpt
Download Desktop · Self-hosting · Features · Tea OpenChart brings AI agents into every step of your investment journey: statistical and technical analysis with code and charts, real-time responses to market events with alerts, opportunity discovery through agentic research, and more. OpenChart is not built to monetize software, but to explore a future where humans and agents manage assets together. If you believe in this future and want to join forces, we want to talk to you. ## Quick start Download Desktop for macOS (Apple Silicon), open the disk image, and install OpenChart. To run Desktop from source, install Node.js 24, npm 11.11.1, Bun 1.4.2, and Just, then run: [code] `just install` also initializes and builds Tea, a programming language built for analyzing market, by the same team. Follow the setup guide for your first chart and alert, or the development guide for building and packaging the app. Desktop is free. AI usage follows your agent provider's plan; optional Cloud market data is priced separately. ## Key features Charts without software quotas Add indicators and Volume Profile, then expand to a 4×4 grid. No software plan limits on charts, indicators, or price and indicator alerts. Use your own agents Use your Claude, Codex, and Gemini models to research markets, build indicators, and set up alerts with the accounts you already have. Describe what to watch Tell your agent what matters. It builds Tea monitors that evaluate incoming data and investigate when a condition fires. Explain what moved the chart Select a range and let your agent research the move, with source-linked annotations pinned to the bars. Draw it, get alerted Draw a line or zone on the chart. When price crosses it, an alert can ask your agent to research the move. Write your own market logic in Tea Build custom indicators and conditions with functions, loops, state, and collections that evaluate as data flows in. Explore Tea → ## Staying updated Star OpenChart to support the project and help more people discover it. ## Credits Thank you to the projects and people whose work helps shape OpenChart: - Jan: for inspiration on the app's design and visual style. - TradingView's Pine Script: for inspiring Tea's language design. - OpenCode: a major inspiration for OpenChart, and the project through which we discovered Effect. - Dify: for inspiring the organization of this README. - assistant-ui and AG-UI: for the agent conversation interface and event protocol. - shadcn/ui, Radix UI, and Base UI: for reusable interface components. - Effect: for the runtime, concurrency, and error-handling foundations. - Inter and Hugeicons: for typography and interface icons. - robust-orientation and related libraries by Mikola Lysenko, and kld-intersections by Kevin Lindsey: for geometry algorithms adapted in Tea. Thanks also to the maintainers of the many other open-source dependencies that make OpenChart possible. ## Contributing Help improve Open