Inference Gateways & Routers

How it works

Gateways put many providers behind one API and one key. Reka is listed as a provider on each of these, so you can reach Reka models and the open source models Reka serves next to everything else you use. Each guide below shows how to call a model and, where the gateway lets you, how to make Reka the provider for a model that several providers serve.

Billing differs by gateway: some bill you for Reka usage themselves, others let you bring your own Reka API key so usage lands on your Reka account at Reka’s rates. Model IDs also differ from Reka’s; each guide lists the ones to use.

• 
  OpenRouterReka models are listed on OpenRouter, so anything that already speaks OpenRouter can route requests to us.
• 
  LLMGatewayReka is a provider on LLMGateway. Pick a Reka model and route through its single endpoint.
• N
  NanoGPTReka is a provider on NanoGPT. Append :reka to a model ID to route it through Reka.
• 
  LiteLLMUse the reka/ provider in the LiteLLM SDK or proxy. Chat completions, Responses and Messages are supported.
• E
  Experiential LabsReka models are in the Experiential Labs catalog. Call them through its OpenAI-compatible gateway.

OpenRouter

OpenRouter lists Reka as a provider for the open models Reka serves (GLM 5.3 Flash, GLM 5.3, DeepSeek V4 Pro), as a Reka endpoint alongside other providers. Requests use your OpenRouter key and are billed by OpenRouter. OpenRouter docs ↗

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "z-ai/glm-5.3-flash",
    "provider": { "only": ["reka"] },
    "messages": [{ "role": "user", "content": "Hello" }]
  }'

LLMGateway

LLMGateway is an open-source, OpenAI-compatible gateway with Reka as the reka-ai provider. It serves glm-5.3 and deepseek-v4-pro through Reka today (GLM 5.3 Flash isn’t routed to Reka there yet), with one API key for every provider. LLMGateway docs ↗

curl https://api.llmgateway.io/v1/chat/completions \
  -H "Authorization: Bearer $LLMGATEWAY_API_KEY" \
  -H "Content-Type: application/json" \
  -H "X-No-Fallback: true" \
  -d '{
    "model": "reka-ai/glm-5.3",
    "messages": [{ "role": "user", "content": "Hello" }]
  }'

NanoGPT

NanoGPT routes to Reka for the open models Reka serves (GLM 5.3 Flash, GLM 5.3, DeepSeek V4, Qwen3.8). Requests use your NanoGPT key and balance. NanoGPT docs ↗

curl https://nano-gpt.com/api/v1/chat/completions \
  -H "Authorization: Bearer $NANOGPT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "z-ai/glm-5.3-flash:reka",
    "messages": [{ "role": "user", "content": "Hello" }]
  }'

LiteLLM

LiteLLM has a native reka/ provider, so the SDK and the proxy call Reka directly with your Reka API key. Chat Completions, Responses and Messages are supported. Requires LiteLLM 1.106 or later. LiteLLM docs ↗

import os
from litellm import completion

os.environ["REKA_API_KEY"] = "your-reka-api-key"

response = completion(
    model="reka/glm5.3-flash",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Experiential Labs

Experiential Labs is an OpenAI-compatible gateway with a provider waterfall per model. GLM 5.3 Flash and the Reka models are in its catalog, callable by slug with an xpl_ key. Experiential Labs docs ↗

curl https://api.experientiallabs.ai/v1/chat/completions \
  -H "Authorization: Bearer $EXPLABS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-5.3-flash",
    "messages": [{ "role": "user", "content": "Hello" }]
  }'

Next steps

• AI Agents & Developer ToolsPower your favorite AI agents and developer tools with Reka’s high-performance inference. Connect directly using your Reka API key.
• ModelsEvery model Reka serves, with context and rates.
• API referenceCall Reka directly: every request and response field.