Supported AI providers
FlowFn integrates with a growing set of AI providers. Each provider is available as a registry in the AI Models section. Connect your API key for a provider and use any of its models in workflow tasks.
Text and structured generation
- OpenAI – GPT models for text generation, structured output, and guardrails. Also supports image generation via DALL-E.
- Google Gemini – Text generation and multimodal capabilities (text, images, audio). Also supports Veo for video generation.
- Anthropic – Claude models for text and structured generation, strong at analysis and reasoning tasks.
- Mistral – Text and structured generation with efficient models for various workloads.
- Cohere – Text generation with models optimized for enterprise use cases.
- Groq – Ultra-fast inference for text generation using hardware-accelerated models.
- Grok (xAI) – xAI's Grok models for text and structured generation.
Image generation
- OpenAI (DALL-E) – Text-to-image generation.
- Stability AI – Text-to-image generation with Stable Diffusion models.
- Fal – Fast text-to-image generation.
- Replicate – Text-to-image generation across various open-source models.
Video generation
- Runway – Text-to-video, image-to-video generation, text-to-speech, and sound effects.
- Google Gemini (Veo) – Text-to-video and image-to-video generation.
Model categories
Models in each registry are tagged by capability to help you choose:
- Reasoning – Best for complex analysis and multi-step problem solving.
- Multimodal – Can process text, images, and other media types.
- Cost-efficient – Good balance of quality and speed at lower token cost.
- Fast – Optimized for low latency and quick responses.
- Creative – Tuned for content creation, writing, and generation tasks.
Choosing a model
When adding an AI Model task to a workflow, select a registry (provider) and then choose a specific model. Consider the task type (text, image, video), required quality, speed, and cost. You can test different models by changing the selection and running the workflow manually.
API key requirements
Each provider requires an API key stored as a named connection on your team. Some plans include default platform keys for select providers. See Using AI Models for setup instructions.
Usage and billing
AI model usage counts toward your plan's credit consumption. Larger prompts, responses, and media generation consume more credits. Check your plan details for monthly caps and top-up options.