Studio (Artifacts)

The Studio panel captures and organizes structured outputs from your Lab conversations. When the AI generates useful artifacts like comparison tables or architecture diagrams, you can save them for reference and export.

Artifact Types

Tables

Comparison matrices, pricing breakdowns, feature lists, and specification tables.

Charts

Benchmark visualizations, cost projections, performance graphs, and trend analysis.

Memos

Executive summaries, decision rationales, recommendation documents, and research notes.

Diagrams

Architecture diagrams, flow charts, system designs, and integration maps.

Automatic Detection

Lattice automatically detects artifacts in AI responses:

| Model | Context Window | Price per 1M tokens |
|-------|----------------|---------------------|
| Claude Sonnet | 200K | $3.00 input / $15.00 output |
| GPT-4 Turbo | 128K | $10.00 input / $30.00 output |
| Gemini Pro | 32K | $0.50 input / $1.50 output |

When a structured element is detected, a Save to Studio button appears.

Manual Saving

You can also manually select any portion of a response and save it:

Highlight the content you want to save
Click Save to Studio in the context menu
Choose the artifact type and add a title
The artifact appears in your Studio panel

Viewing Artifacts

Click any artifact in the Studio panel to:

View full content — See the complete artifact
Copy to clipboard — Quick export for documents
Edit title — Rename for better organization
Delete — Remove artifacts you no longer need

Artifact Properties

Each artifact stores:

Property	Description
`title`	User-defined name for the artifact
`type`	table, chart, memo, or diagram
`content`	The artifact data (markdown, JSON, or SVG)
`source_message_id`	Link to originating conversation
`created_at`	Timestamp of creation

Use Cases

Comparison Tables

Generate side-by-side comparisons for decision-making:

"Create a comparison table of Claude, GPT-4, and Gemini
for production RAG applications. Include context window,
pricing, latency, and key limitations."

Executive Memos

Generate board-ready documentation:

"Write an executive memo recommending our AI model choice
for the Q1 chatbot launch. Include cost projections and
risk assessment."

Architecture Diagrams

Visualize system designs:

"Create an architecture diagram showing how our RAG pipeline
should integrate with Claude API, including caching and
fallback providers."

Benchmark Charts

Visualize performance data:

"Create a chart comparing latency benchmarks for Claude Haiku,
GPT-4 Mini, and Gemini Flash at different request volumes."

API Reference

List Artifacts

GET /api/workspaces/{workspace_id}/artifacts

Create Artifact

POST /api/workspaces/{workspace_id}/artifacts
Content-Type: application/json

{
  "title": "Model Comparison Q1 2024",
  "type": "table",
  "content": "| Model | Context | Price |..."
}

Get Artifact

GET /api/workspaces/{workspace_id}/artifacts/{artifact_id}

Update Artifact

PATCH /api/workspaces/{workspace_id}/artifacts/{artifact_id}
Content-Type: application/json

{
  "title": "Updated Title"
}

Delete Artifact

DELETE /api/workspaces/{workspace_id}/artifacts/{artifact_id}