Introducing eroq — expressive AI in one credit-metered API

For two years we have been building consumer character products: chat that stays in persona for hundreds of messages, galleries rendered overnight, voice lines that ship with the reply. The models, the sampling, the retry logic, the pricing meter — all of it exists because our own products beat on it daily.

Today that stack becomes an API. eroq exposes ten models behind one key:

• RP+ (eroq-rp-plus) — our flagship roleplay model. 3 credits per completion.
• RP mini (eroq-rp-mini) — the high-volume tier. 1 credit per completion.
• Krea2 (eroq-krea2) — clean photoreal image generation. 10 credits.
• Eroq Uncensored (eroq-uncensored) — uncensored photoreal. 10 credits.
• Eroq One (eroq-one) — native 2K generation and instruction editing with up to 10 reference images. 20 credits.
• Anima (eroq-anime) — anime illustration, uncensored. 10 credits.
• Motion One (eroq-motion-one) — our own video engine: 5–10s cinematic clips with native audio, uncensored. 60 credits per clip.
• Voice One (eroq-voice-one) — studio-grade TTS. 3 credits per 100 characters.
• Voice Turbo (eroq-voice-turbo) — low-latency speech. 2 credits per 100 characters.
• Scribe One (eroq-scribe-one) — transcription. 1 credit per minute of audio.

Why credits

Per-token pricing is precise and hostile. Nobody planning a product knows what "$2.40 per million output tokens" means for their unit economics until they have shipped, measured, and been surprised.

So we price the way we wished someone had priced for us: every call has a flat cost in credits. A completion is 1 or 3 credits whether the scene ran long or short. An image is 10 credits — 20 on the 2K reference engine — whether the prompt was one line or ten. You can compute your margin on a napkin before writing a line of code — and a failed generation refunds itself, because paying for nothing is absurd.

Credits are prepaid, start at $10 for 1,000, and never expire.

Why "expressive"

General-purpose assistants are tuned to be helpful, brief, and careful. Character products need the opposite: models that commit to a persona, write with texture, and do not break the fourth wall to apologize. RP+ is tuned for exactly that — the tuning is its own story.

The same philosophy runs through the visual and audio models: they are chosen and configured for creative range. Dark, violent or intense fiction renders instead of hitting a refusal wall, inside the hard limits of our acceptable use policy.

The parts we sweated

The shape is boring on purpose. Requests follow the OpenAI convention where one exists. Most client code ports by changing the base URL, the key, and the model id.

Docs cannot drift. Reference pages, openapi.json and llms.txt are generated from the same spec object the endpoints charge from. If the docs say 3 credits, the meter says 3 credits, because they are the same line of code.

Streaming is first-class. stream: true gives you standard SSE chunks, and a final usage event carries the meter before [DONE].

Your outputs are yours. Generated media comes back inline in the response, and every render is also saved to your own Library with the full recipe that made it — free, private until you publish it, and deletable whenever you want. Pass private: true on the image and video endpoints and not even that is kept: no file, no library entry, and the prompt shows as stars in the usage ledger.

FAQ

What models does eroq run today?

RP+ and RP mini for chat; Image One, Image Anime and Image Art at 10 credits each; Voice One at 3 credits per 100 characters and Voice Turbo at 2; Scribe One at 1 credit per minute of audio. On video, Motion One is the uncensored engine at 60 credits for 5 seconds, alongside Seedance 2.5, Seedance 1.0 Pro and Lite, Kling 2.5 Turbo, Hailuo 02 and Veo 3 Fast, which are moderated upstream by their providers and gated by plan.

How much do credits cost?

Packs start at $10 for 1,000 credits, with $50 for 5,500 and $100 for 12,000. New accounts get 50 free credits, nothing expires, and failed generations refund themselves. Monthly plans grant credits that roll over forever — the whole grid is on /pricing.

Is the API OpenAI-compatible?

Where a convention exists, yes. chat/completions takes a messages array and streams standard SSE chunks, images/generations and audio/speech follow suit, and most client code ports by changing the base URL to https://eroq.ai/v1, the key and the model id. Video generation is async by necessity — POST a job, then poll or take a signed webhook.

Is there an interface, or is it API only?

Both. The same engines drive a studio you sign into — video, image, voice and text, with a persistent cast of characters — and anything the studio does, it does by calling the same public API you have.

Start now

Sign up, mint a key, and the quickstart has you at a first completion in one curl. New accounts carry 50 free credits — enough to try every model before a card ever enters the picture.

We built this because we needed it. If you are building a character product, you probably do too.