API direct · Official API resourcesAvailable in API direct mode
Your models. Your resources. Your choice.
One place for AI
Choose your access
Connect models to your apps and creative workflows; Choose official API resources with API direct, or explore a range of model resources through a subscription plan
API direct · Pay as you goOfficial API resources
Subscription plansModel resources tailored to your plan
Different sources, features, and usage limits; Compare before you connect
API direct; Your routing rulesExample
API direct · Official API resources
Routing group0.50×1.20×
View current routes & prices1 / 3
A0.82×Eligible
B0.95×Over limit
C1.05×Over limit
What does the multiplier mean?
A route at 0.80× costs $0.80 for usage valued at $1 at the model’s reference rate. These illustrative rates are not live quotes.
You pay the matched rate, not your ceiling.1 example route fits your limit. Availability is checked when you call.
Your providers; Your price ceilingPay the matched route’s customer rate
Know what you connect toSource and access method vary by mode
Access options
Choose the access that fits your work
Compare resources and billing at a glance
Official API resources
API direct
Pay as you go from your prepaid balance
- Choose a provider and routing group
- Pay the matched route’s rate
- Set your price ceiling
Resource details
Official API resources come from your selected provider. Your price ceiling filters eligible routes; usage is charged at the matched route’s rate. Check model features and route availability before connecting.
Platform plan resources
Subscription plans
Model access with plan allowances
- Model resources included in your plan
- Access differs from official APIs
- Plan allowance and prepaid balance stay separate
Resource details
Plans include platform-integrated access and platform-hosted open-source models. Resources and access methods vary by plan. Review supported features, usage limits, and data handling before subscribing.
API protocol adaptation, built in
One Key.Agents meet models.
Pick your tools and models. We handle the protocols.
Keep your agent’s API protocol and reach different models with the same key. Tokatlas bridges protocol differences, so you can skip building your own adapter.
Terminal agents
Editor agents
Custom agents
Tokatlas protocol adaptation
One API key. Multiple API protocols.
- OpenAI
- Anthropic
- Gemini
Multiple AI models
Model APIs connected. Responses in your agent’s format.
See the integration docs for supported protocols, model combinations, and tool-calling capabilities.
Built around your work
Different goals. One place to connect.
Find the right starting point for what you want to do.
From API key to working application.
Connect official model API resources to your product through pay-as-you-go. Give development and production their own keys, choose resources for each workload, and inspect requests when something goes wrong.
Read the integration docs- 01Choose your modelReview capabilities and API compatibility
- 02Configure your keyPayment mode, resources, and price ceiling
- 03Inspect your requestsRequest logs and recorded costs
Model discovery
Start with what you want to make.
Explore text, image, and video models for your application or tools.
Compare capabilities and billing units.
OpenAIGPT-5
Explore this model for text-based applications and developer workflows.
Text modelView model
GoogleGemini 2.5 Flash Image
Explore image generation for your creative integrations and compatible tools.
Image modelView model
Video generationSeedance 2.0 Mini
Explore video generation and check the task format before connecting.
Video modelView model
Explore each model’s access options, plan availability, supported features, and billing units in the catalog and plan details.
From choice to first request
A clear path to getting started
01 / DISCOVER
Find a model for your task
Check capabilities, API compatibility, and pricing for the models you want to use
02 / CONFIGURE
Create a key with your settings
Choose a payment source; for pay-as-you-go, select resources and set your price ceiling
03 / CONNECT
Connect, then inspect the result
Use the API or a compatible tool, then check request logs and recorded costs in your account
Before you begin
A few things worth knowing
How do API direct and subscription resources differ?
API direct uses official model API resources. Subscription plans combine platform-integrated access, including subscription- or client-based integrations, with platform-hosted open-source models. These access methods differ from official APIs. Your purchase covers the platform plan rather than a model provider’s own subscription; see each plan for resource details.
Do I pay my price ceiling on every request?
No. For pay-as-you-go keys, you pay the matched route’s customer rate. The ceiling sets the maximum rate you are willing to accept; it is not the price charged for every request.
What is the difference between Reliability and Value?
Reliability prioritizes available resources in the stability-focused group. Value prioritizes lower-priced resources in the value-focused group. Both pay-as-you-go groups use official API resources. The choice changes routing priorities, not whether the resource is official. Performance can still vary by route.
Can I use Tokatlas without writing code?
You can connect through a third-party tool that supports the relevant API and custom credentials. Check the tool’s compatibility first. Tokatlas’s model API is not itself an image or video editor.
Are prepaid funds and plan allowances interchangeable?
They are separate payment resources. Your key’s payment mode determines how usage is billed. Subscription model access and allowances follow the selected plan.
Does a successful video request mean my video is ready?
Not necessarily. Submitting a request and completing generation are different stages. Check the video task’s status and the related cost records.
Your next idea starts with a connection.
Choose your models. Set your controls. Start building.