Creates a new message in a specific AI thread.
Permission | Description |
---|---|
ai_thread_messages.create | Create messages in your own thread |
Name | Type | Required | Description | Example |
---|---|---|---|---|
thread-id | int | Yes | ID of the thread | 123 |
Name | Type | Required | Description | Example |
---|---|---|---|---|
content | string | Yes | Content of the message (max 5000 chars) | "Hello!" |
state | object | No | Optional state data as JSON | { ... } |
{
"content": "Hello!",
"state": { "form": { "field": "value" } }
}
Returns the created message and the AI's response message.
{
"createdMessage": { /* ai thread message resource */ },
"responseMessage": { /* ai thread message resource */ }
}
For a full schema, see AI Thread Message Resource.
In
responseMessage.updated_state
you may find the updated state data you initially sent.
Status | Description | Reference |
---|---|---|
422 | Validation error | Validation error |
401 | Unauthorized | Authentication error |
403 | Forbidden (no permission) | Permission error |
429 | Too many requests (rate limit exceeded) |
The AI system uses OpenAI APIs behind the scenes, it handles reading Dambel documentation files, searching web or reading a specific web page, saving memories for each user and access them in other chats, reading data from Dambel's database, and even reporting the conversation whenever necessary. Once it uses every tool that it has, it returns the final response to the message.