by Unity Nodes
docs.near.aicloud › experimental › direct completions

NEAR AI Cloud: Direct Completions

Updated 02-11-2026

View on docs.near.ai

Experimental: send requests straight to one self-hosted model's TEE endpoint, without the NEAR AI Cloud Gateway.

[!WARNING]

Experimental. Direct completions endpoints are not recommended for new integrations or production verification workflows. Use the NEAR AI Cloud Gateway (https://cloud-api.near.ai/v1) instead.

Current verification limitations:

  • An attestation report covers one instance. Several instances can serve the same direct hostname, and separate connections can reach different ones. GET /v1/attestation/report returns the report of the instance that answered, and all_attestations contains only that report. You can't verify every serving instance before a completion, so a TLS key pinned from one report fails when a later connection reaches another instance, even a legitimate one. Tracked in cloud-api#1087.
  • Signature lookups can return 404. Each instance keeps response signatures in its own memory. GET /v1/signature/{chat_id} can reach a different instance than the one that served the completion, and there is no stable way to choose the instance.

A direct completions endpoint sends your request straight to one model's TEE at its own hostname, without the Gateway in between. It accepts the same API key as the Gateway. Use it only when you maintain an existing direct integration.

Endpoint format

https://{slug}.completions.near.ai/v1

Each hostname serves one model. The endpoint registry lists every hostname and the model it serves:

curl -s https://completions.near.ai/endpoints | jq '.endpoints[]'

{ "domain": "glm-5-3-flash.completions.near.ai", "models": ["z-ai/glm-5.3-flash"] }

The registry can list endpoints for models that aren't in the public catalog. Only build on an endpoint whose model appears on the Models page.

Send a request

curl

curl https://glm-5-3-flash.completions.near.ai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "z-ai/glm-5.3-flash",
    "messages": [{"role": "user", "content": "Hello, NEAR AI!"}]
  }'

Python

from openai import OpenAI

client = OpenAI(
    base_url="https://glm-5-3-flash.completions.near.ai/v1",
    api_key="YOUR_API_KEY",
)

response = client.chat.completions.create(
    model="z-ai/glm-5.3-flash",
    messages=[{"role": "user", "content": "Hello, NEAR AI!"}],
)

print(response.choices[0].message.content)

Routes

Each hostname exposes the routes its model supports:

  • POST /v1/chat/completions: chat completions
  • POST /v1/completions: text completions
  • POST /v1/tokenize: tokenization
  • POST /v1/embeddings: embeddings
  • POST /v1/rerank: reranking
  • POST /v1/score: scoring
  • POST /v1/images/generations: image generation
  • POST /v1/images/edits: image editing
  • POST /v1/audio/transcriptions: audio transcription
  • POST /v1/privacy/classify: privacy classification, on privacy-filter endpoints
  • GET /v1/models: the model served by this hostname
  • GET /v1/attestation/report: attestation report for the instance that answers
  • GET /v1/signature/{chat_id}: response signature, if the answering instance has it

Verify a direct endpoint

Direct verification is experimental. Each check applies only to the request or connection you inspect. It doesn't establish evidence for later requests to the same hostname.

Move to the Gateway

To switch an existing direct integration to the Gateway:

  1. Change the base URL to https://cloud-api.near.ai/v1.
  2. Set model to the Gateway model ID from GET https://cloud-api.near.ai/v1/model/list, for example z-ai/glm-5.3-flash.
  3. Keep your API key. It works on both.

Then follow Verification to check Gateway requests.

Back to Documentation