Skip to content

Prompts

A prompt is a stored text snippet you reuse — opening hooks, briefs, style guides, structured generation requests. Prompts themselves are not "run" by a dedicated endpoint; they're seeds you reference from AI Posts and the composer.

Endpoints

GET    /api/v1/ai-prompts            list (filters: search, team_id)
GET    /api/v1/ai-prompts/:id        fetch one
POST   /api/v1/ai-prompts            create   { prompt, team_id? }
PUT    /api/v1/ai-prompts/:id        update
DELETE /api/v1/ai-prompts/:id        delete

Model shape

typescript
{
  user_id: ObjectId,    // creator
  team_id?: ObjectId,   // optional team association
  prompt: string,       // template body
  // timestamps: createdAt, updatedAt
}

Variables in the body (, ) are conventional — there is no server-side templating; the frontend prompts you for each placeholder before the model call.

Running a prompt

To execute a prompt against a model, the frontend calls POST /api/v1/ai-generate/text (the same endpoint as Inline) with the resolved prompt body and an optional modelId. AI Posts (campaign-style bulk generation) also reads from the prompt library.