AI providers (Flow AI Gateway)
Connect your own API keys, set default models, and choose which provider each Studio feature uses.
Open AI providers in the Studio#Gateway dashboard metrics
The top of the Flow AI Gateway displays live summary metrics across your workspace: Active Providers, default models for Text, Visual, and Audio (Voice Engine), and total recorded Gateway Calls.
#My AI connections
Together AI offers its three strongest models in each kind: text (Qwen3.5 9B, Qwen3.7 Plus, MiniMax M3), images (Nano Banana 2, Nano Banana Pro, FLUX.2 Pro — all take reference images) and video (Wan 2.7 text-to-video, Wan 2.7 image-to-video, Seedance 2.5). AI33.pro / OpenSpeaker is used for audio only (voices and narration in Audio Studio); it is not offered for images or video. A model you had saved that is no longer offered switches to that provider's default automatically.
Add an API key per provider (for example Gemini, OpenAI, Anthropic, DeepSeek, Together AI, AI33/OpenSpeaker). Keys are encrypted on the server and never shown again after saving. Your own keys are used before any platform key the administrator allows.
Test connection / Check provider connections confirms a key works.
#My AI defaults and per-task routing
- My AI defaults: The provider and model per generation type (text, image, audio, video).
- Personal Studio AI routing: A clean per-task routing table to customize providers for specific Studio stages: Script Outline, Script Polish, Scene Breakdown, Narration Audio, Visual Generation, Thumbnail Design, Topic Research, and Competitor Analysis.
- Effective configuration: Shows the order Flow will try providers in, falling back gracefully if your preferred provider faces temporary rate-limits.
#YouTube Data API
Topic research, competitor tracking and YouTube thumbnail references use the YouTube Data API. Create a Google Cloud project, enable YouTube Data API v3, create an API key and paste it here.