Router Setup

Use Wafer through Vercel AI Gateway, OpenRouter, Bifrost, and TrueFoundry AI Gateway.

Use these setup paths when your app already sends traffic through a model router or gateway. The router owns the API key, base URL, model slug, billing, and routing policy; Wafer serves the underlying model capacity.

For direct Wafer Serverless API usage, use API Reference. For coding agents and local harnesses, use Agent Setup.

Vercel

Wafer is available through Vercel AI Gateway for GLM 5.2 Fast.

Use the Vercel AI SDK model slug:

import { streamText } from "ai";

const result = streamText({
  model: "zai/glm-5.2-fast",
  prompt: "Why is the sky blue?",
});

See Vercel’s GLM 5.2 Fast via Wafer changelog for provider-specific details.

OpenRouter

Wafer is listed as a provider on OpenRouter.

Use OpenRouter when you want Wafer models behind OpenRouter’s unified API key, catalog, usage tracking, and routing controls.

OpenRouter model slugs are model-level slugs, not provider-specific slugs. To prefer Wafer for GLM 5.3, use z-ai/glm-5.3 and put the Wafer provider first in provider.order (Wafer is listed as the wafer endpoint on that model page). Fallbacks remain enabled, so OpenRouter can route to another provider if Wafer is unavailable.

curl -sS "https://openrouter.ai/api/v1/chat/completions" \
  -H "Authorization: Bearer <YOUR_OPENROUTER_API_KEY>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "z-ai/glm-5.3",
    "provider": {
      "order": ["wafer"]
    },
    "messages": [
      {"role": "user", "content": "Reply with the single word: ready."}
    ]
  }'

See OpenRouter’s Quickstart for the API shape and Provider Routing for provider selection controls.

Bifrost

Wafer is an OpenAI-compatible provider with a dedicated provider implementation in Bifrost.

Use Bifrost to configure Wafer provider keys and model allowlists, then call Wafer through Bifrost’s OpenAI-compatible endpoints.

1. Add Wafer as a provider

   In Bifrost, go to Models → Model Providers. Look for Wafer under Configured Providers. If it is missing, click Add New Provider and select Wafer.

   Click Add Key or edit an existing key. Set a name, then paste your Wafer API key or reference an environment variable such as env.WAFER_API_KEY. Set Allowed Models to All Models (the default) or to the specific model allowlist you want the key to serve.

   To upload files, enable Use for Batch APIs on the key. This flag also gates file operations. Save the provider configuration.

2. Call Wafer through Bifrost

   Send requests to Bifrost’s OpenAI-compatible Chat Completions endpoint and prefix the Wafer model ID with wafer/. Bifrost serves every model Wafer exposes; call GET /v1/models for the current catalog.

   curl -X POST "http://localhost:8080/v1/chat/completions" \
     -H "Content-Type: application/json" \
     -d '{
       "model": "wafer/DeepSeek-V4-Flash-0731-Fast",
       "messages": [{"role": "user", "content": "Hello!"}]
     }'
   

File operations reuse Bifrost’s batch key filter. Enable Use for Batch APIs on at least one Wafer key, or uploads fail with no config found for batch apis.

See Bifrost’s Wafer provider docs for supported operations and request parameters.

TrueFoundry

TrueFoundry AI Gateway is the proxy layer that sits between your applications and the LLM providers and MCP Servers. It is an enterprise-grade platform that enables users to access 1000+ LLMs using a unified interface while taking care of observability and governance.

Wafer is available natively in TrueFoundry AI Gateway. See TrueFoundry’s Wafer provider docs for the source setup flow.

1. Add a Wafer account

   In TrueFoundry, go to AI Gateway → Models, select Wafer, and add your Wafer account.

   The default Wafer base URL is https://pass.wafer.ai/v1. Change it only if Wafer directs you to a different endpoint. Add your Wafer API key for authentication. Enable Zero Data Retention in TrueFoundry when you want it to send Wafer-ZDR: required on requests.

   Add collaborators to the account if other users or teams should be able to access it.

2. Register models

   Click + Add Models and register the Wafer model IDs you want TrueFoundry to expose. For each model, set a display name, the exact Wafer model ID, and select Chat as the model type.

   Common Wafer model IDs include GLM-5.3, Kimi-K3, and Qwen3.5-397B-A17B.

3. Call the gateway

   Use the TrueFoundry model ID format your-wafer-account/model-display-name through TrueFoundry’s OpenAI-compatible /chat/completions API:

   from openai import OpenAI
   
   client = OpenAI(
       api_key="your_truefoundry_api_key",
       base_url="{GATEWAY_BASE_URL}",
   )
   
   response = client.chat.completions.create(
       model="my-wafer-account/GLM-5.3",
       messages=[{"role": "user", "content": "Hello, how are you?"}],
   )
   
   print(response.choices[0].message.content)