Skip to content
TIE Docs

Overview

TIE is a platform of two services that work together:

TIE Auth — Authentication service that handles user registration, login (email/password + OAuth), and token management. Built on Google Identity Platform / Firebase.

TIE AI Gateway — OpenAI-compatible API that sits between your application and LLM providers. Adds persistent memory, personas, observability, and client-side tool calling — while exposing the same /v1/chat/completions endpoint that every OpenAI SDK already knows.

How They Work Together

Section titled “How They Work Together”
flowchart LR
    App["Your App"]

    subgraph TIE["TIE Platform"]
        Auth["Auth"]
        Gateway["AI Gateway"]
    end

    App -- "1. Sign up or log in" --> Auth
    Auth -- "2. Get back a token" --> App
    App -- "3. Send messages" --> Gateway
    Gateway -- "4. AI response" --> App

    style App fill:#f9f9f9,stroke:#d1d1d1,color:#181818
    style Auth fill:#fff7ed,stroke:#F68220,color:#181818
    style Gateway fill:#fff7ed,stroke:#F68220,color:#181818
    style TIE fill:#fff7ed08,stroke:#F68220,color:#181818

That's it — 4 steps:

  1. Get a token — Call TIE Auth to sign up or log in. You get back a Bearer token.
  2. Use the token — Pass it as Authorization: Bearer <token> on every AI request.
  3. Send messages — POST /v1/chat/completions with your messages, just like OpenAI.
  4. Get AI responses — TIE handles memory, personas, and model routing automatically. You just get the response back.

TIE Auth Features

Section titled “TIE Auth Features”
  • OAuth sign-in — Google and Apple, via client-side token flow
  • Email/password — Registration and login
  • Token management — JWT tokens with 1-hour expiry, refresh tokens for re-authentication
  • User profiles — Metadata, roles, account suspension

TIE AI Gateway Features

Section titled “TIE AI Gateway Features”
  • OpenAI-compatible — Drop-in replacement for OpenAI, Cloudflare AI Gateway, or LiteLLM
  • Custom prompt mode — Send your own system or developer message to override the agent's built-in prompt while keeping memory and personas
  • Memory — Persistent user memory via Graphiti knowledge graph, injected automatically
  • Personas — Custom system prompts per user, injected automatically
  • Client tool calling — Pass tool definitions per-request, TIE routes tool_calls back for execution
  • Internal tools — Memory search/write execute server-side, invisible to the client
  • Thread persistence — Maintain conversation state across requests
  • Audio — Speech-to-text and text-to-speech via OpenAI-compatible endpoints
  • Image Generation — Generate images via OpenAI-compatible /v1/images/generations with OpenAI and Google Imagen models
  • Observability — All requests traced via Langfuse
  • Multi-agent — Select different agents via X-Agent-Id header

Supported Providers

Section titled “Supported Providers”
  • Anthropic (via Vertex AI) — Claude Haiku 4.5, Claude Sonnet 4.5, Claude Sonnet 4.6, Claude Opus 4.5, Claude Opus 4.6
  • OpenAI — GPT-5 Nano, GPT-5 Mini, GPT-5.1
  • Google Vertex AI — Gemini 2.5 Flash, Gemini 2.5 Pro, Gemini 3 Flash

Query GET /info on your instance to see which models are currently available.

Documentation

Section titled “Documentation”
PageDescription
AuthenticationSign up, log in, OAuth, token management, and refresh flows
Chat CompletionsSend messages via the OpenAI-compatible /v1/chat/completions endpoint
AudioSpeech-to-text and text-to-speech via /v1/audio/transcriptions and /v1/audio/speech
Image GenerationGenerate images via /v1/images/generations with OpenAI and Google Imagen
AgentsMulti-agent routing, agent selection, and the X-Agent-Id header
ThreadsConversation threads — list, rename, delete, and history
MemoryPersistent memory, context prefetch, and memory management
PersonasCustom system prompts per user — create, activate, and manage