RequestsOne API · your provider keys0% markup
ClaudeWrite the reply1.1s
GPTInvestigate the bug2.4s
GeminiClassify the ticket0.3sModel gateway
Any model through one API
Switch models in configuration instead of in your agent’s code. Use your own provider keys at no markup, or let Adapt hold the keys.
ClaudeWrite the reply1.1sGPTInvestigate the bug2.4sGeminiClassify the ticket0.3sUse cases
What you can build with model gateway.
Switch models without rewrites
Change the model in configuration instead of in your agent’s code.
Bring your own keys
Keep your provider accounts and their prices, with no markup from Adapt.
A model for each step
Send routine classification and in-depth investigation to different models.
How it works
Use different models for different parts of a workflow.
Choose the tradeoffs for your application instead of sending every task to the same model.
Define the tasks
Separate routine classification from a more involved investigation or writing task.
Choose the requirements
Pick the model for each task based on quality, response time and cost.
Evaluate and adjust
Review results on your own workload and change the configuration as your requirements evolve.
import { Adapt } from '@adapt/sdk'
const reply = await new Adapt().models.chat({
model: 'anthropic/claude-sonnet',
keys: 'yours',
messages: [{ role: 'user', content: prompt }],
})Pricing
What it costs to run.
Use your own provider keys at no markup, or let Adapt hold the keys and bill the models with your usage.
- 0%markup on your own provider keys
- +5.5%on provider list price through Adapt’s keys
Works with
Use it with the rest of Adapt.
The company agent combines all four with your tools, shared knowledge and the places your team already works.