Skip to content
@maximhq

Maxim AI

Enterprise AI Gateway to Route, Govern, and Secure all your AI traffic
image

At Maxim AI, we're building the AI infrastructure for enterprises to route, govern, and secure all their AI traffic.

AI runs across a company's applications and employee devices. Each model request, tool call, or agent action raises questions: what data is being shared, who has access, and which policies apply? We built Bifrost to give enterprises visibility and control across all AI traffic with the performance and reliability their production workloads require.

Bifrost Gateway routes every model request, tool call, and agent action through shared security and governance controls. Teams can route AI traffic intelligently across multiple models and providers, gain visibility into the AI traffic, control org-wide access and budgets for AI tools and models through their existing identity structures set on tools like Entra and Okta, apply a unified set of security policies and guardrails, and optimize costs.

Bifrost Edge extends that protection to employee devices, where AI tools can otherwise operate outside the company's controls. It is a lightweight service that runs on those devices and routes AI traffic from the tools people already use through the Bifrost Gateway, so those tools come under the same policy as everything else.

This is what closes the shadow AI gap. Desktop chat apps, AI in the browser, coding agents in the terminal and editor, and the MCP servers those tools connect to would otherwise never reach the Gateway, and most organizations have no visibility into them at all. Edge also makes the experience better for the people using those tools. There is one sign-in through the company's existing identity provider, no keys to paste into local configuration files, and no new steps to learn. Only AI traffic is routed, and nothing else on the device is inspected.

Bifrost runs air-gapped in your infrastructure, with no data, not even telemetry, coming to us. Its Apache 2.0 core is free to run and contribute to; the enterprise tier adds advanced security, identity-based governance, high availability, and support.

Explore Bifrost

Quick start | Website | Documentation | Discord | Enterprise | Bifrost LLM Gateway | Bifrost MCP Gateway | Bifrost Edge | Changelogs

Pinned Loading

  1. bifrost bifrost Public

    Open-source, high-performance AI gateway that works as your LLM gateway, MCP gateway and agent gateway. One OpenAI-compatible API for 10,000+ models, with failover, load balancing, semantic caching…

    Go 8.8k 1.4k

  2. maxim-cookbooks maxim-cookbooks Public

    Maxim is an end-to-end AI evaluation and observability platform that empowers modern AI teams to ship agents with quality, reliability, and speed.

    Jupyter Notebook 17 8

  3. maxim-go maxim-go Public

    SDK to integrate Maxim in your Go app.

    Go 9 3

  4. maxim-js maxim-js Public

    Maxim's Typescript/Javascript SDK

    TypeScript 10 3

  5. maxim-py maxim-py Public

    Maxim's Python SDK

    Python 11 2

  6. maxim-java maxim-java Public

    Public issue tracker for Maxim java SDK

    Kotlin 1 1

Repositories

Showing 10 of 33 repositories
  • bifrost Public

    Open-source, high-performance AI gateway that works as your LLM gateway, MCP gateway and agent gateway. One OpenAI-compatible API for 10,000+ models, with failover, load balancing, semantic caching, virtual keys, budgets and MCP Code Mode. Works with any SDK or agent and adds under 200µs of overhead at 5k RPS.

    maximhq/bifrost's past year of commit activity
    Go 8,757 Apache-2.0 1,385 544 (20 issues need help) 664 Updated Oct 13, 2026
  • bifrost-benchmarking Public

    Reproducible AI gateway benchmarks: Bifrost vs LiteLLM and Portkey on latency percentiles, throughput and memory, using Vegeta and a mock LLM provider. Includes a load generator and the MCP Code Mode benchmark. Published runs: 9.5x throughput and 54x lower P99 than LiteLLM.

    maximhq/bifrost-benchmarking's past year of commit activity
    Go 12 Apache-2.0 7 0 12 Updated Oct 12, 2026
  • maxim-py Public

    Maxim's Python SDK

    maximhq/maxim-py's past year of commit activity
    Python 11 Apache-2.0 2 1 16 Updated Oct 5, 2026
  • maxim-cookbooks Public

    Maxim is an end-to-end AI evaluation and observability platform that empowers modern AI teams to ship agents with quality, reliability, and speed.

    maximhq/maxim-cookbooks's past year of commit activity
    Jupyter Notebook 17 8 0 60 Updated Oct 5, 2026
  • maxim-go Public

    SDK to integrate Maxim in your Go app.

    maximhq/maxim-go's past year of commit activity
    Go 9 Apache-2.0 3 0 2 Updated Sep 30, 2026
  • maxim-js Public

    Maxim's Typescript/Javascript SDK

    maximhq/maxim-js's past year of commit activity
    TypeScript 10 Apache-2.0 3 10 (1 issue needs help) 3 Updated Sep 25, 2026
  • bullfrog-agent Public Forked from bullfrogsec/agent
    maximhq/bullfrog-agent's past year of commit activity
    Go 0 MIT 3 0 0 Updated Sep 24, 2026
  • .github Public
    maximhq/.github's past year of commit activity
    0 1 0 0 Updated Sep 21, 2026
  • actions Public

    Collection of Maxim Github Actions

    maximhq/actions's past year of commit activity
    JavaScript 0 1 0 7 Updated Sep 21, 2026
  • vt10x Public Forked from hinshun/vt10x

    Package vt10x is a vt10x terminal emulation backend.

    maximhq/vt10x's past year of commit activity
    Go 0 MIT 49 0 1 Updated Sep 2, 2026