TokLab
TokLab is one API for AI image, AI video, LLM and AI audio models. See every price before you call, and pay only for the requests that succeed. No subscription.
What is TokLab?
TokLab is a single API that gives developers access to AI image, video, LLM and audio models from providers such as OpenAI, Anthropic, Google, xAI, Meta, ByteDance, MiniMax, DeepSeek, Kimi and Kuaishou. It is aimed at developers and teams who call more than one model and want one key, one balance and one way to read results. What sets TokLab apart is its billing rule: failed or rejected requests cost nothing, and credits never expire, with no subscription to cancel.
Key features
- One key, every model
- Image, video, language and audio models on TokLab share one API key, one balance and one way to read results.
- Failed requests cost $0
- A timeout, a filtered prompt or an empty render is not billed, so retries and fallbacks never cost twice.
- Credits never expire
- The balance has no expiry date and no monthly reset, and there is no subscription to cancel.
- Prices visible before you call
- Every model lists its live price, its parameters and the request format it accepts, so the cost and shape of a call are known in advance.
- Familiar request formats
- TokLab speaks the OpenAI, Anthropic and Gemini formats, plus one task API for media, so moving to another model means changing one field rather than adding an SDK.
- One task API for media
- Image, video and audio jobs return a task id immediately; you poll it and download the output file from TokLab storage.
- Consistent error codes
- The same failure returns the same code on every model, such as 30001 for adding credits, 30002 for backing off and retrying, and 60001 for revising the prompt.
- Auditable usage
- Every request lands in the TokLab usage history with its charge, alongside task logs and LLM logs showing tokens in and out.
How to use TokLab
- Create an account by signing in to TokLab with Google. Every model and its price are visible before you add any money.
- Add credits by topping up in USD with a card, PayPal or crypto.
- Create an API key in the TokLab dashboard; you can make one key per app and revoke any of them.
- Point your client at the TokLab base URL (
https://api.toklab.ai/v1), set your key and pick a model. The Quickstart has a working example for each request type, including OpenAI SDK, Claude Code, Codex, video, image and audio.
Use cases
- Developers who want to call image, video, LLM and audio models from several providers through one integration and one balance.
- Teams running agents such as Claude Code or Codex on TokLab models, or connecting an MCP client to list models, read prices and submit jobs.
- Media pipelines that generate video from a prompt or a still image, generate or edit images from prompts and reference images, and download the finished files from TokLab storage.
- Workloads that transcribe or translate speech with Whisper-1 and need JSON, SRT or VTT output.
- Builders who retry heavily or experiment with prompts and want failed attempts to cost nothing.
Pricing
TokLab has no subscription and bills per request, with failed requests charged at $0.00 and credits that never expire. You top up in USD by card, PayPal or crypto. Live per-model rates shown on the site include:
- Seedance 2.5: $0.10262 per second
- GPT Image 2.5: $0.00448 per image
- GPT Image 2: $0.00425 per image
- Midjourney V8.2: $0.08 per job
- Nano Banana 2.1: $0.021 per image
- PixVerse V6: $0.02 per second
- Kling 4.0: $0.1425 per second
- Whisper-1: $0.036 per minute
- Claude Opus 5.5: $1.20 per 1M input tokens, $6.00 per 1M output tokens, $0.06 per 1M cache read, $1.50 per 1M cache write
- GPT-6 Sol: $0.20 per 1M input tokens, $1.00 per 1M output tokens
- Gemini 3.1 Pro: $0.80 per 1M input tokens, $4.80 per 1M output tokens
Some models come in more than one price tier, which you name in the model field. Check the official website for the full and current rate list.
Is TokLab free to use?
TokLab has no free plan or subscription. You add credits in USD and are charged per request, but failed or rejected tasks cost $0.00, so you only pay for requests that succeed.
Which models can I call through TokLab?
TokLab lists 44 models across four categories: 16 video models, 9 image models, 18 LLM models and 1 audio model. Providers include OpenAI, xAI, Google, Anthropic, ByteDance, MiniMax, DeepSeek, Kimi, Meta and Kuaishou.
Do TokLab credits expire?
No. TokLab states that your balance has no expiry date and no monthly reset, so you can top up once and spend it this week or next year.
Can I keep using the OpenAI SDK with TokLab?
Yes. TokLab is OpenAI-compatible and also accepts Anthropic Messages and Gemini request formats, plus a task API for image, video and audio jobs. You change the base URL and the API key rather than adding a new SDK.
What happens on TokLab if a request fails?
A failed task is not billed and costs $0.00. TokLab also returns the same failure code on every model, each with a suggested next step, so your code can branch on the error.
How is TokLab different from calling each model provider directly?
TokLab gives you one key, one balance and one billing rule across providers, with each model's live price, parameters and request format listed before you call. Unlike per-attempt billing on most platforms, failed tasks are free and there is no subscription to manage.
TokLab
TokLab is one API for AI image, AI video, LLM and AI audio models. See every price before you call, and pay only for the requests that succeed. No subscription.
Cloudflare
HTTP/3
Next.js
Node.js
React
WebpackFree AI anime and manga generator. Turn text into anime episodes, manga, comics, webtoons and comic shorts in any art style — no drawing or animation skills needed.
Translate image text across 70+ languages with our advanced AI Image Translator to help you better expand your products globally to various countries
Advertised Here
Reach thousands of visitors daily. Get your spot now!
Tools similar to TokLab
Simplify your prompt management: consolidate, test, and track AI prompts all in one spot, making collaboration and version control a breeze.
Combining cutting edge developer tools & AI Agents to enable autonomous software development - SuperAGI
No-code platform for creating an AI web app that serves your audience just like you do. No technical skills required.
LiteLLM handles loadbalancing, fallbacks and spend tracking across 100+ LLMs. all in the OpenAI format
Create Chatbots from Postgres, Websites, Google Sheets, Notion, Files, Text, HTML, and more.
Typeblock is a no-code tool to build, share, and monetize AI apps in minutes
The platform for AI phone calling
