GPT Proto puts more than 200 models from different vendors behind one OpenAI-compatible API and one account balance. Text covers GPT 5.6, Claude Opus 5 and Gemini 3.7 Flash; images run through Seedream 5.0 Pro, GPT Image 2 and Nano Banana; video through Kling v3, Dreamina Seedance and Vidu Q3, with editing tools alongside. Smart routing picks the cheapest available channel automatically.
Multi-key budget control is the feature that makes this usable by a team rather than an individual. Shared API credit is a standing problem – one person’s runaway loop drains the balance everyone depends on, and provider consoles rarely offer per-key caps. Issuing keys with their own ceilings turns a shared account into something a finance function can live with.
Rates are published per model rather than hidden behind credits: Claude Opus 5 at $4 per million input and $20 output, Gemini 3.7 Flash at $0.45 and $2.25, Seedream 5.0 Pro at $0.0405 a run, Kling v3 4K at $0.336 a second. There is free fallback routing with no platform fee and recharge bonuses that scale with top-up size. An aggregator inserts a third party between you and every provider’s terms and uptime, cheapest-channel routing means you do not always know which channel served a request, and no company entity is named anywhere.
