OpenAI Assistant
Connect your virtual being to an OpenAI Assistant through the OpenAI Assistants API.
The OpenAI Assistant integration connects your virtual being to an assistant created in the OpenAI Platform. Unlike the standard OpenAI AI Model integration (which uses the Chat Completions API), the OpenAI Assistant engine uses the Assistants API, which supports persistent conversation threads, file search (RAG over uploaded files), function calling, and code interpretation.
Prerequisites
- An OpenAI account with API access
- An Assistant created in the OpenAI Platform (platform.openai.com → Assistants)
- The Assistant's ID and your OpenAI API key
Configuration
Credentials
| Field | Description |
|---|---|
| API key | Your OpenAI API key |
| Assistant ID | The unique identifier of the Assistant you want to connect. Found in the OpenAI Platform under Assistants – it looks like asst_XXXXXXXXXX |
Finding Your Assistant ID
- Go to platform.openai.com/assistants
- Click on the Assistant you want to use
- Copy the Assistant ID from the assistant details panel (format:
asst_...)
How It Works
When the Call Conv AI node routes to the OpenAI Assistant engine:
- Virbe creates or retrieves a persistent Thread for the current conversation session
- The user's message is added to the thread as a new message
- An Run is created for the thread using your configured Assistant
- OpenAI processes the run, executing any tools or file searches the Assistant is configured with
- The Assistant's response is returned to Virbe and delivered to the user
Thread continuity is maintained for the lifetime of the Virbe conversation session.
Comparison with AI Model (Chat Completions)
| Feature | AI Model (Chat Completions) | OpenAI Assistant |
|---|---|---|
| Conversation history | Managed by Virbe | Managed by OpenAI (Threads) |
| File search / RAG | Via Virbe Knowledge Base | Via OpenAI file_search tool |
| Function calling | Via node logic | Via Assistant tool definitions |
| Code interpreter | No | Yes (if enabled on Assistant) |
| Best for | Standard LLM responses | Complex agents with OpenAI-native tools |
Use Cases
OpenAI Assistant is a strong choice when you:
- Have an existing OpenAI Assistant already configured with tools, files, or custom instructions
- Need OpenAI's native file search over uploaded documents (PDF, Word, etc.)
- Want to use the code interpreter for calculations or data analysis
- Prefer OpenAI-managed conversation threading
The OpenAI Assistants API may have higher latency than Chat Completions due to the asynchronous run model. For latency-sensitive deployments, consider using the standard AI Model configuration with the Tool Agent node instead.