NEAR AI Launched VSCode Extension for GitHub Copilot
The integration enables private AI inference via hardware-secured enclaves to protect sensitive developer codebases.
Updated on Oct. 1, 2026 in Artificial Intelligence
Live Poll
Would you switch your coding tools to prioritize data privacy over mainstream model providers?
NEAR AI Cloud has released a new VSCode extension allowing users to connect GitHub Copilot to private, hardware-secured inference environments. This integration provides a pathway for developers to keep source code inaccessible to infrastructure providers during AI-assisted workflows.
Why it matters
As enterprises grow wary of exposing proprietary code to large model providers, this tool uses hardware-level verification to ensure data privacy without sacrificing compatibility. It marks an effort to decentralize AI infrastructure by allowing users to stake tokens for service credits.
The platform supports 258 OpenAI-compatible options and uses hardware-signed cryptographic attestations to verify that inference jobs run in secure environments. Input costs range from $0.15 to $0.50 per million tokens.
The players
NEAR AI Cloud
A cloud platform providing inference services and decentralized compute access.
GitHub Copilot
An AI-powered coding assistant platform integrated into development environments.
VSCode
A popular open-source code editor and extension platform developed by Microsoft.
The details
The VSCode extension functions by utilizing an OpenAI-compatible API to route traffic from Copilot Chat to the NEAR AI Cloud. It incorporates hardware-secured enclaves—a set of protected physical memory partitions that ensure data is isolated from the main processor and other software—to maintain privacy. Users can also enable client-side encryption, which renders inputs unreadable to the NEAR infrastructure itself during the execution process.
Timeline
December 2025: NEAR AI Cloud launched.
June 2026: VS Code update enabled BYOK support.
August 2026: Platform integrated SayGm gateway.
September 2026: OpenRouter integration occurred.
October 1, 2026: Official news publication.
The Tech Race
The integration follows the trajectory set by hardware-secured enclaves, applying the security model to decentralized AI inference. It competes with centralized LLM provider privacy offerings by leveraging cryptographic attestations rather than relying on trust-based service agreements.
Developers can now direct Copilot traffic to private instances by configuring the extension to use their own NEAR-staked credits. The workflow change requires no shift in existing IDE habits, but necessitates an active NEAR wallet to provision inference credits.
The takeaway
Developers concerned about code privacy should monitor how these cryptographic attestations scale across different model architectures. The move indicates a shift toward verifiable compute, where trust is replaced by hardware-signed execution logs.
Further reading
For broader context on the evolution of private model hosting, see our Artificial Intelligence coverage.
Source note: This article includes information reported by Crypto Briefing.
Live Poll
Would you switch your coding tools to prioritize data privacy over mainstream model providers?
