Start
Start
Documentation Menu

MCP Server

Dal Nulla exposes a Model Context Protocol (MCP) server that lets you generate images, videos, and text directly from AI coding assistants like Claude Code, Cursor, and Windsurf.

Quick Setup

Choose your client and copy the configuration:

Terminal 
claude mcp add dalnulla --transport http --url https://mcp.dalnulla.com/mcp

How to get your token: Go to your dashboard, open the browser DevTools, and run JSON.parse(Object.entries(localStorage).find(([k]) => k.includes('auth-token'))?.[1] ?? '{}').access_token in the console. Copy the result as your bearer token.

Authentication

The MCP server uses Bearer token authentication via Supabase. Every request must include a valid access token in the Authorization header. Tokens are obtained from Supabase Auth and are the same tokens used by the Dal Nulla web app.

Available Tools

The MCP server exposes 10 tools organized into three categories:

Generation

generate_image

Generate images from text prompts using 23+ models (FLUX, Midjourney, OpenAI, Runway, etc.)

generate_video

Generate videos from text/image prompts using 20+ models (Veo, Kling, Sora, Runway, etc.)

generate_text

Generate text using LLMs (Gemini, GPT, Grok) with optional multimodal input

enhance_image

Upscale images (up to 4x) or remove backgrounds using Topaz and BiRefNet models

Utility

list_models

List available models by type (image, video, text) with free/pro status

get_credits

Check your credit balance (used, remaining, max) for the current billing period

estimate_cost

Estimate the credit cost of a generation before running it

Workflows

list_projects

List your workflow projects with IDs (needed for run_workflow)

run_workflow

Execute a complete workflow project end-to-end and return results

get_workflow_runs

List recent workflow executions with status and media counts

Tool Reference

generate_image

Generate images using AI models (FLUX, Midjourney, OpenAI, Runway, etc.)

Parameter Type Required Description
prompt string Yes Description of the image to generate
model string No Model key
Default: nano-banana
aspectRatio string No Aspect ratio
Default: 1:1
Values: "1:1", "16:9", "9:16", "4:3", "3:4"
resolution string No Image resolution
Default: 1K
Values: "1K", "2K", "4K"
count number No Number of images (1-4)
Default: 1
negativePrompt string No What to avoid in the image
seed number No Seed for reproducibility
sourceImageUrls string[] No Source image URLs for image-to-image generation

Usage Examples

Once configured, you can use Dal Nulla tools directly in your AI assistant. Here are some example prompts:

Generate an image 
Use the generate_image tool to create a "cyberpunk cityscape at sunset, neon lights reflecting on wet streets" in 16:9 aspect ratio at 2K resolution
Generate a video from an image 
Use the generate_video tool with the image I just generated as startFrameUrl to create a 10-second cinematic pan across the scene
Check costs before generating 
Use estimate_cost to check how much a 4K image with flux-2-pro would cost, then generate it if I have enough credits
Run a workflow 
List my projects with list_projects, then run the workflow for the "Product Showcase" project using run_workflow

Free Models

The following models are available on the free plan:

  • Image: nano-banana, nano-banana-pro
  • Video: veo-3.1-generate-preview, veo-3.1-fast-generate-preview
  • Text: gemini-3-flash-preview, gemini-3-pro-preview, gemini-2.5-pro

All other models require a Pro subscription.