OpenAI
GPT Image
Complex, instruction-heavy scenes and iterative edits in one session.
GPT Image handles narrative prompts with many constraints — “left third empty for copy,” specific objects, layered instructions — and iterative refinement without rebuilding the whole prompt from scratch.
What you get
- Generated media from this model family
What you can send
EchoClip matches Generate workflows and the model picker to these slots — not a single upstream default model.
Required (at least one route)
- Source image + edit instruction (refine)
- Text prompt
- Video clip
Optional
- Layout constraints in prompt
- Multiple references
- Strength / steps controls
- Style / ratio options
- Text prompts
Studio routes in this family
Each fal endpoint has its own input contract. Pick the route that matches the media you have (start frame, FLF, references, audio, etc.).
Example workflows
- Mixed Media — Video clip · Optional: Text prompts · Style / ratio options · Strength / steps controls
When to use it
- Complex scenes with multiple subjects and layout rules
- Iterative edits from a previous generation in the same thread
- Concept art when the brief reads like a creative director note
Reach for something else when
- You need maximum photoreal product fidelity — compare Flux
- Strict brand hex enforcement is mandatory — pair with palette tools in Generate
EchoClip MCP (Cursor & Claude)
Connect EchoClip in chat to quote credits, pick this model family, and kick off renders without leaving your editor.
Long briefs work well in MCP chat — agent can create_image, critique, and re-run with edits.
MCP setup guides — authenticate once, then generate from chat.
Workflows in EchoClip
- Complex prompts
- Iterative refine
- Concept boards
Best for
- Concept art
- Layout-heavy stills
- Instruction-dense briefs
Use GPT Image in Generate
Open the Image studio, pick this model in the picker, and check live credit estimates before you render.