OpenChart: Open-Source TradingView Alternative with Your Own AI Agent
OpenChart is an open-source charting workspace that plugs your existing Claude, Codex, or Gemini plan directly into the chart. Instead of driving markets through a CLI, agents annotate setups, write indicators, investigate moves, and create alerts on the same canvas you're looking at.
Key details
- Local-first: charts, research, and indicator code stay on your machine. The backend runs locally too.
- Bring your own agent plan: works with Claude, Codex, and Gemini accounts you already pay for. AI usage follows your provider's plan — no software cap on charts, indicators, or alerts.
- Free market data from day one, with an optional paid cloud data feed for lower latency.
- Charting: add indicators and Volume Profile, expand to a 4×4 grid. No plan limits on chart count or alert count.
- Draw-to-alert: draw any line or zone on a chart and attach an alert. When price crosses it, the alert can trigger your agent to investigate the move and save the research into the workspace.
- Range research: select a price range and the agent investigates the move, returning source-linked annotations pinned to the bars.
- Agent monitors: describe what you care about, and the agent builds Tea monitors that evaluate incoming data and investigate when conditions fire.
Tea, the market DSL
Custom indicators and alert conditions are written in Tea, a programming language by the same team. It supports functions, loops, state, and collections that evaluate as data flows in — same idea as Pine Script, which the README credits as an inspiration. Running just install initializes and builds Tea as part of setup.
Install and run
macOS on Apple Silicon only right now. Download the disk image and install, or build from source with Node.js 24, npm 11.11.1, Bun 1.4.2, and Just:
git clone https://github.com/longsurf-ai/openchart.git
cd openchart
just install --frozen-lockfile
just desktopRun just check before submitting PRs. Development and setup guides are in docs/; the runtime is built on Effect, with UI from shadcn/ui, Radix, and Base UI.
Who it's for
Developers who already have an AI coding plan and want market tooling that lives on their own box with agent-written indicators and alerts — not a hosted SaaS chart.
📖 Read the full source: HN LLM Tools
👀 See Also
Selfware: Rust-based local AI agent framework with PDVR architecture
Selfware is an open-source AI agent framework built in Rust for local inference, implementing a PDVR cognitive cycle with 54 built-in tools and designed for long-running tasks on consumer hardware.
Xiaozhen: A Claude Code skill that digs three layers into root causes
Xiaozhen (小真) is a Claude Code skill that uses three mechanics—The Gift, Three Layers Deep, and The Prediction—to help users uncover what's actually bothering them rather than giving direct advice. It's installed with a one-line curl command and activated by typing /小真 in Claude Code.
The Companies Cutting Headcount for AI Will Lose to the Ones Who Didn't
AI headcount cuts trade long-term institutional knowledge for short-term savings. Retaining teams and using AI to amplify their capability is the winning strategy.
Design Studio Plugin for Claude Code Adds Virtual Design Team with 9 Roles and 16 Commands
A new Claude Code plugin called Design Studio simulates a full design team with 9 specialist roles, 16 slash commands, and 5 agents. It auto-detects tech stacks and includes over 8,000 lines of design knowledge across reference files.