Hedra
Log inSign Up
Open Hedra
A woman in a cobalt velvet suit floating above a leather sofa in a red velvet screening room, a grey tabby cat watching from the cushion below
AI image generator

The AI image generator you can just talk to.

Tell Hedra's agent what you need. It writes the prompt, picks the right model, and reworks the result with you until it's right.

Ask the agent

You'll create an account to run it — the conversation carries over

What people are making

All made here · Hover for recipes
Barista laughing with a regular at a sunlit Roman espresso bar“Roman espresso bar, early morning, candid laughter”SEEDREAM 5.0 PRO · 2K — RUN →Woman flipping through a record crate whose hand-lettered dividers read Jazz, Soul and Disco“Record shop crate — dividers read "Jazz / Soul / Disco"”GPT IMAGE 2 HIGH · 2K — RUN →Man in a powder-blue suit loading plush carnival animals into a cream sedan outside a donut shop“Plush prizes, cream sedan, roadside donut shop”SEEDREAM 5.0 PRO · 2K — RUN →
Fisherwoman in a conical hat on a teal boat in misty Ha Long Bay at dawn“Fisherwoman at dawn, Ha Long Bay”SEEDREAM 5.0 PRO · 2K — RUN →Skier in a mustard jacket carving through deep powder against a cobalt sky“Powder skier, mustard shell, cobalt sky”SEEDREAM 5.0 PRO · 2K — RUN →
Your image goes here.Generate Image
Ripe purple figs on a dew-covered branch in golden backlight, one split open“Ripe figs on the branch, painterly, golden light”SEEDREAM 5.0 PRO · 2K — RUN →
Anime key visual of a girl in a red scarf on a rooftop at dusk with a crow on the railing“Anime key visual, rooftop at dusk, red scarf”SEEDREAM 5.0 PRO · 2K — RUN →Runners crossing a marathon finish line beneath a FINISH banner and a race clock reading 2:41:07“Marathon finish — clock reads "2:41:07"”GPT IMAGE 2 HIGH · 2K — RUN →Night-market chef flambéing a wok under neon signs, flame and vegetables frozen mid-air“Night-market wok flambé, neon, frozen mid-toss”NANO BANANA PRO · 1K — RUN →

Get the details right.

Woman with copper-red curls in a cobalt velvet suit reclining on a mustard sofa with three grey tabby catsSEEDREAM 5.0 PRO · CHARACTER REFERENCEThe same face, every imageUpload a reference once — the character holds across every scene you generate.
Corner diner at dusk with a neon sign reading OPEN ALL NIGHT and window lettering reading COFFEE, PIE, RECORDSGPT IMAGE 2 · NATIVE TEXTText that spells — signs and labels render correctly.
Hands with chrome-yellow nails opening a cobalt-blue soda can labeled Fizz and Co Blue Raspberry SodaNANO BANANA PRO · TARGETED EDITEdit one detail — the rest stays exactly as it was.
Ultra-wide rainforest canopy after rain with layered palm fronds and hanging mangoesNANO BANANA PRO · 4K NATIVENative 4K — full resolution out of the model, no upscaler pass, no artifacts.

What the agent handles.

Writes the prompt
Your one sentence becomes a model-grade prompt — lighting, lens, composition, style references — without you learning prompt syntax.
Picks the model
Photorealism, illustration, readable text, design assets — each job goes to the model that's best at it. Override any time.
Keeps characters consistent
Reference a face or product once — the agent carries it across every image in the session.
Iterates in plain language
"Moodier." "More room for the headline." "Same scene, winter." Feedback is the interface — no re-prompting from scratch.
The agent picks the right one for the job

Every state-of-the-art image model, under one agent.

GoogleGoogleNano Banana ProNano Banana 2Imagen 4
OpenAIOpenAIGPT Image 2 HighGPT Image 2 MediumGPT Image 2 Low
Black Forest LabsBlack Forest LabsFlux.2 [max]Flux.2 [pro]Flux.2 [flex]Flux.2 [klein] 9BFlux Kontext MaxFlux Kontext Pro
ByteDanceByteDanceSeedream 5.0 ProSeedream 5.0 Lite
xAIxAIGrok Imagine
See all models →

Tell it what you're making.

Ask the agent

You'll create an account to run it — the conversation carries over

FAQs

One image. Infinite possibilities.

Victorian glass greenhouse at dawn with sun rays through fogged panes
Bioluminescent jellyfish glowing violet in deep indigo water

Start generating.

Ask the agent

Every image around this box was made on this page's models