Skip to main content

​
POST /api/suggest

Generate intelligent inline suggestions based on the current context. This endpoint integrates with the user’s memory system to provide personalized, context-aware completions.

​
Request Body

​
context
string
required
The current text context where the suggestion will be inserted. This can be selected text or the text around the cursor.
​
userId
string
required
Unique identifier for the user. Used to retrieve personalized memories.
​
instruction
string
Optional instruction to guide the suggestion. For example: “complete this sentence” or “suggest next steps”.
​
model
string
default:"gpt-4o-mini"
The AI model to use for generation. Supported models include:
  • gpt-4o-mini (default)
  • gpt-4o
  • claude-3-5-sonnet-20241022
  • groq/llama-3.3-70b-versatile
  • cerebras/llama-3.3-70b

​
Response

The endpoint returns a streaming text response with the AI-generated suggestion.
​
suggestion
string (streaming)
The generated text suggestion, streamed token-by-token for real-time display.

​
Memory Integration

The suggest endpoint uses three memory tools to provide personalized suggestions:
  1. searchMemory - Searches relevant memories based on the context
  2. addMemory - Stores new facts learned during the interaction
  3. getAllMemories - Retrieves all user memories for comprehensive context
The system automatically:
  • Searches for relevant personal preferences and past context
  • Incorporates your coding style, preferences, and recent work
  • Stores new insights for future suggestions

​
Example Request

const response = await fetch('http://localhost:3001/api/suggest', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    context: 'const handleSubmit = async () => {',
    userId: 'user_123',
    instruction: 'Complete this function',
    model: 'gpt-4o-mini'
  })
});

const reader = response.body?.getReader();
const decoder = new TextDecoder();

// Read streaming response
while (true) {
  const { done, value } = await reader.read();
  if (done) break;
  
  const chunk = decoder.decode(value);
  console.log(chunk); // Display suggestion as it streams
}

​
Example Response (Streaming)

  try {
    const formData = new FormData();
    formData.append('data', JSON.stringify(data));
    
    const response = await api.post('/submit', formData);
    if (response.ok) {
      showSuccessToast('Submitted successfully');
    }
  } catch (error) {
    showErrorToast(error.message);
  }
}

​
Use Cases

Code Completion

Intelligent code completions that match your style and patterns

Writing Assistance

Context-aware text completions for emails and documents

Next-Step Suggestions

Suggest logical next steps based on current context

Personalized Help

Suggestions that incorporate your preferences and history

​
Keyboard Shortcut

Trigger inline suggestions with Ctrl+Space in the Tabby desktop app. Source: nextjs-backend/src/app/api/suggest/route.ts
  • Completion - General text transformations and completions
  • Chat - Full conversational AI with memory
  • Search Memory - Directly query the memory system

Build docs developers (and LLMs) love

Get started for freeTalk to us