NEAR AI Cloud

What it does

NEAR AI Cloud is a hosted inference API for open-weight language models. Requests go through an OpenAI-compatible endpoint, and each model runs inside a trusted execution environment built on Intel TDX with NVIDIA confidential GPUs. Every response carries a hardware-signed attestation that binds the code and weights that served it, and a model catalog lists per-token rates for each model.

Developers point existing OpenAI SDK code at the NEAR AI base URL with an API key, then optionally verify attestation and response signatures using the documented verification tooling. The runtime, verifier and enclave images are described as open source and reproducible. A separate NEAR AI Red program pairs abliterated cybersecurity models with confidential inference and a first-month discount for security teams.

It suits developers and teams handling sensitive data who want private inference without rebuilding their stack, and enterprises that need reserved capacity or private models through a quoted agreement. Keys are created in the Cloud console at cloud.near.ai. Pay-as-you-go has no minimums or subscriptions and is billed per token by model, while enterprise terms are quoted. Verification tooling differs between NEAR-hosted and third-party-hosted TEE tiers.

What it can do

• Run open-weight models inside Intel TDX and NVIDIA confidential-GPU enclaves
• Return a hardware-signed attestation with each response for verification
• Serve an OpenAI-compatible API that works with existing OpenAI SDK code
• Provide a model catalog with per-token input and output pricing
• Offer per-key API access with spend caps and per-key usage reporting
• Support enterprise options such as reserved GPU capacity and private models

Works with

• OpenAI SDK (OpenAI-compatible base URL)

Before you choose

• Pay-as-you-go is billed per token by model, with input and output rated separately, so costs vary by model choice.
• Private TEE and Private 3P TEE tiers differ in where the enclave runs and which verification tools apply, so check the model's documentation.
• Enterprise features such as reserved capacity, private models, and invoicing are quoted and require contacting the team.

Listed under

• Coding
• Confidential computing
• Developers
• LLM API
• Open-weight models
• Private inference
• Verifiable AI

Get started

Sign in to the Cloud console and create a key; no new SDK needed.

1. Sign in to the NEAR AI Cloud console at cloud.near.ai and create an API key.
2. Install the OpenAI Python SDK and set base_url to https://cloud-api.near.ai/v1 with your key in NEAR_AI_KEY.
3. Call a listed model ID such as Qwen/Qwen3.8-27B, then optionally verify attestation using the documentation at docs.near.ai/cloud/verification.

Billed per 1M input and output tokens at each model's rate. No platform fee on top of the model rate; attestation and TEE execution are included. Free allowance and trial terms were not confirmed. Enterprise pricing is quoted separately.

Source: near.ai/pricing, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Private AI inference you can verify | NEAR AI ↗near.ai
• Private. Intelligent. Yours. - NEAR AI ↗docs.near.ai
• Per-token pricing for private AI inference | NEAR AI ↗near.ai
• Cybersecurity Models and NEAR AI Red Launch — NEAR AI ↗near.ai
• NEAR AI + Levanto Labs: Verifiably Private Decision Models — NEAR AI ↗near.ai

Company website

near.ai: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

manual_review

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.