Docs · Questions
ChatTDUsage
What are all the ChatTD release versions and features?
ChatTD Release History
ChatTD is a multi-API chat operator for TouchDesigner with UI and multimodal support. 2 major releases.
v2.0.0 - March 24, 2024
Major rebuild with UI and expanded features.
Multi-API Support
- Over 120 models available total
- OpenRouter.ai - Multiple providers
- OpenAI - Official API
- Groq - Fast inference
- LM Studio - Local inference
- Easy API server selection and key entry
- “Find API Key” button for quick access
- API keys stored locally in Pythonvenv folder
New TD UI
Fancy TouchDesigner-native interface built with new UI techniques.
Stream Toggle
- Per-token response streaming
- Messages come back in chunks
- ‘Interrupt Stream’ toggle to stop early
Multimodal Image Integration
- Include images in conversation
- Multiple images per message supported
- Drag and drop images onto operator
- TOP reference support
- ‘Include Chat Images’ toggle for all images in conversation
Automatic Conversation Management
- Auto-saves in JSON within folder structure
- Create, save, and load conversations
- Dropdown menu for conversation selection
Other Features
- Parameter UI reorganized/cleaned
- ‘Stop Phrases’, ‘Seed’, ‘Frequency Penalty’, ‘Presence Penalty’ in Settings 2
- Callbacks: onGenerate, onDone, onError
- Easy Python environment setup (requires Python 3.11)
- Token count and cost monitoring (estimated, non-streaming only)
- Extension class for advanced control
Availability
DotExplorer tier
v1.0.0 - October 24, 2023
Initial release - ChatGPT API operator
Features
- Callable extension class
- Callbacks: onGenerate, onDone, onError
- ‘Update/Add to Chat’ toggle to freeze conversation
- Conversation saving and loading
- Table-based full conversation (editable/customizable)
- Outputs: last/selected response, CHOPs, full conversation, chat history
- Quick conversation edits (clear table, remove last message)
Latest Version: v2.0.0 (March 24, 2024)
Total Releases: 2
Supported APIs: OpenRouter, OpenAI, Groq, LM Studio (120+ models)
Requirements: Python 3.11