GPT-5.5
Reasoning-tuned flagship. Best for complex orchestration and tool use across the canvas.
Traycer is a multi-model LLM proxy framework for real-time collaborative vector graphics. Orchestrate GPT-5.5, Claude 4.6, and Gemini 3.1 on a single infinite canvas.

The Proxy
Route every prompt to the model that fits. Hot-swap mid-stroke with smooth, sub-100ms transitions.
Reasoning-tuned flagship. Best for complex orchestration and tool use across the canvas.
Long-form precision. Surgical edits, refactors, and document-grade vector annotation.
Multimodal native. Reads diagrams, screenshots, and SVG fragments as first-class inputs.

The Canvas
Pen, frame, arrow — meet prompt. Traycer's toolbar treats AI as a first-class drawing primitive.
Live cursors, presence, and conflict-free vector ops.
Prompt to layout. Refine with natural language.
Production-ready, accessible SVG with semantic IDs.
Round-trip diagrams as code, instantly.
Connectivity
Auto-commit generated components and SVG assets to your repo.
Sync diagrams as embedded, live-updating canvases.
Send traced flows as inline outreach assets.
Two-way frame sync.
Drop interactive previews.
Hydrate diagrams from schema.
Credits
No seats. No monthly minimums. A single transparent credit balance across every model.
Balance
Top up
Credits never expire. Volume rates kick in past 1M.