Photo Goblin

loading your darkroom…

◆ Photo

New adjustment layer

Theme

AI: not configured

Layers

Click any entry to go back to that point.

Sign in to continue

Your edits stay where they are — signing in does not touch the picture you have open. For the agent you bring your own OpenAI or Gemini API key; it is stored against your account and never sent back to the browser.

By continuing you agree to the Terms of Service and Privacy Policy.

AI settings

AI provider

OpenAI

API key

Agent model

Vision model

Image model

Speech-to-text model

Text-to-speech model

Gemini

API key

Agent model

Vision model

Image model

Speech-to-text model

Text-to-speech model

Custom

Any OpenAI-compatible chat API — a self-hosted model, OpenRouter, Groq, and the like. Used for the agent and vision analysis only; image generation, transcription and speech stay on OpenAI or Gemini.

Name

Base URL

API key

Agent model

Vision model

Accepts images (vision)

Set for you: Test connection sends one pixel and records the answer, and a run that gets refused turns this off by itself. Off, the agent skips visual tools. Text-only endpoints — DeepSeek's chat models among them — reject any request carrying a picture.

Show the model's reasoning

When there is more than one answer Offer me the options Use your best judgement

Adjust

Logs

↑↓ move Enter run Esc close