Skip to main content
Screenshot of the TrueFoundry Welcome Page with signup option

Welcome Page

  • Click on Signup button.
Screenshot of the TrueFoundry Signup Page for account creation

Signup Page

  • Create an account and verify your email. You will be redirected to the landing page.
Screenshot of the TrueFoundry Landing Page after signup

Landing Page

  • Click on Setup AI Gateway card.

Add your models

Add your provider accounts, API credentials, and models

​
Select your Provider

Go to the AI Gateway Tab from the left sidebar and click on Models tab and select your provider from the list. You can add multiple API keys from the same provider by creating multiple provider accounts.
AI Gateway Models Page showing provider selection

AI Gateway Page

​
Add Provider Account and Models

A Provider Account refers to one account of a model provider like OpenAI, Anthropic, Bedrock (AWS Account) etc.
You can add multiple OpenAI, AWS, Anthropic accounts each with their own API keys / auth tokens as different provider accounts. Each provider account can have multiple models that are available with that provider. For example, if the provider account type is OpenAI, you will have models like GPT-4, GPT-4o, GPT-3.5-turbo etc.
Provider Account Setup Page for adding model providers

Provider Account Page

  • Fill the form and select the model from the list of checkboxes.
In case the model you want to add is not available in the list, you can still add the models by clicking on +Add Model at the bottom of the form and filling up the details.
  • Fill the form and click on Add Provider Account.
  • On submitting, you will see the provider accounts added and the models in those provider accounts in the Models tab in Gateway.
To add additional models in the future, you can do so by traversing to AI Gateway tab and then Models tab. You will see list of provider accounts on the left side, select the provider account and then click on Add Model button on top right corner.Add Models to Provider Account interface in TrueFoundry
The LLM Playground is a UI for the AI Gateway where you can try out different models you’ve added from across providers like OpenAI, Mistral, Cohere etc. Below is an overview of the features:
AI Gateway Playground Page for testing LLM models

AI Gateway Playground Page

  • You can click on Try in Playground button to try out the model in the playground. (if you are on the listing page)
Select Model in Playground interface

AI Gateway Select Model Page

You can access your models via code snippets.
  • Click on Code Snippets tab from the playground page. It will generate the code snippet for the model you have selected.
Gateway Code Snippets Page with code generation options

AI Gateway Code Snippets Page

  • This provides code snippets in multiple languages and frameworks such as OpenAI client, Langchain Client, Python (Streaming and Non-Streaming), Node.js, LangGraph, LLamaIndex, Google ADK, Curl, and more.
  • Additionally, it gives you the option to easily copy your API Key, Model ID, and Base URL for seamless integration.
Integrate Gateway Models in Code example page

Integrate Gateway Models In Code

​
Frequently Asked Questions: Troubleshooting

Verify Proxy API and Headers.Learn more: Proxy API | Headers
Check ratelimiting policy scope (user/key/model).Learn more: Rate Limiting
Validate budgetlimiting rules & cost tracking.Learn more: Budget Limiting | Cost Tracking
Confirm provider credentials and model IDs on the provider page (e.g., Anthropic/Bedrock/Azure).Learn more: Supported Providers
Enable streaming, adjust parameters, consider caching/rerank, use load-balancing to route to lower latency models, review analytics.Learn more: Chat Completions Overview | Load Balancing Overview | Analytics
Refine guardrail config or use custom guardrails.Learn more: Guardrails Configuration | Custom Guardrails
If you have any questions, please reach out to us at support@truefoundry.com or book a demo with our team here.