Bearer authentication header of the form `Bearer <token>`, where `<token>` is your auth token.

Bearer Auth

Prompt version

Get prompt versions

Get a single prompt version

Authorization

Docs

Documentation

API reference

Integrations

Changelog ↗

Get started

Discord

Keywords AI is a full-stack LLM engineering platform for developers and PMs.

What is Keywords AI?

Log your LLM requests and responses asynchronously

AI observability

Set up production monitoring

Keywords AI LLM proxy supports you call 250+ LLMs using the same input/output format.

LLM proxy

Prototype with an AI gateway

Managing prompts is a crucial part of building AI applications. Migrate your prompts to Keywords AI to manage them more effectively.

Migrate your prompts

A guide on how to create and versiona prompt in Keywords AI.

Create a prompt

Deploy a prompt

Monitor a prompt in production and get detailed metrics.

Monitor a prompt

Test and compare prompts for different models in the LLM playground

Improve a prompt

Improve prompts with the model playground

Share a prompt

Overview

Call 200+ LLMs with a single OpenAI compatible format

Integrate 200+ LLMs

This is the guide for how to add your own credentials in Keywords AI

Add own credentials

This is the guide for how to set up fallbacks in Keywords AI

Add fallback models

Increase your LLM rate limits with our load balancing feature.

Enable load balancing

Reduce latency and save LLM costs by caching LLM prompts and responses.

Cache LLM responses

Prompt caching is a feature that allows you to cache the results of a prompt so that it can be reused later.

Enable prompt caching

When an LLM call fails, our system detects the error and retries the request to prevent failovers.

Configure retries

This is the guide for how to use function calling in Keywords AI

Function calling

How to create your own golden datasets from logs.

Create a dataset from logs

Easily manage and organize test cases. Import a CSV file and edit it like a Google Sheet.

Import dynamic test cases

A spreadsheet-like interface built for multi-config prompt testing

Compare & debug prompts

API keys management

Learn how to create a new team in Keywords AI.

Create a new team

Use your own API keys through Keywords AI

Learn how to use OpenAI SDK with Keywords AI LLM proxy.

OpenAI SDK

Anthropic SDK

LangChain SDK

Vercel AI SDK

LlamaIndex

BAML is a domain-specific language to write and test LLM functions. Use BAML with Keywords AI to get complete LLM observability

BAML

Mem0 provides a memory layer for LLM applications, enabling them to remember and learn from user interactions over time.

Mem0

OpenAI

Use your own Anthropic credits through Keywords AI

Anthropic

Use your own Azure OpenAI credits through Keywords AI

Azure OpenAI

Use your own Vertex AI credits through Keywords AI

Google Vertex AI

Use your own AWS credits through Keywords AI

AWS Bedrock

Use your own Groq credits through Keywords AI

Groq

Use your own Fireworks credits through Keywords AI

Fireworks

Use your own Together AI credits through Keywords AI

Together AI

Use your own Perplexity AI credits through Keywords AI

Perplexity AI

Use your own OpenRouter credits through Keywords AI

OpenRouter

PostHog is an open-source product analytics platform, you can monitor your product traffic and user sessions with PostHog. Now you can connect PostHog with Keywords AI to monitor your LLM performance and get observability.

PostHog

Logging API

LLM Proxy

Batch logging

Log list

Update logs

Prompt

Text to speech

Speech to text

Embeddings

User creation

User list

User detail

User update

Evaluations

Model list

Errors

Discover the importance of LLM monitoring in AI. Learn how it ensures accuracy, performance, and reliability while reducing costs. Keywords AI offers top-tier LLM monitoring tools.

What is LLM monitoring?

Retry failed request due to busy upstream provider

Integration methods

Logs

Prompt

Multimodal integrations

User

Evaluations

Model

Prompt version

Get prompt versions

Get a single prompt version

Integration methods

Logs

Prompt

Multimodal integrations

User

Evaluations

Model

​Get prompt versions

​Get a single prompt version

Get prompt versions

Get a single prompt version