⚠️ DEPRECATED - NEAR AI Agent Framework > [!WARNING] > This project is deprecated. Server components will be shut down on October 31, 2025. Deprecation FAQ Why deprecate NEAR AI…
⚠️ DEPRECATED - NEAR AI Agent Framework
> [!WARNING]
> This project is deprecated. Server components will be shut down on October 31, 2025.
Deprecation FAQ
Why deprecate NEAR AI Agent Framework and Developer Hub?
As AI agents have become more popular and more advanced, they've also become more integrated into 'regular' software. There's not that much need (or demand) anymore for a special hosted framework for chat-based AI agents. More often, we're seeing agentic workflows integrated into new or existing applications. For that, existing frameworks like LangChain, AutoGen, and many others are more than sufficient.
What's NEAR AI doing?
NEAR AI has built a new platform for private & confidential AI, consisting of:
- NEAR AI Private Chat (https://private.near.ai/)
- NEAR AI Cloud (https://cloud.near.ai/)
What will happen to the NEAR AI Developer Hub?
The NEAR AI Developer Hub at https://app.near.ai/ will shut down on October 31, 2025 and replaced with this notice. After that date, all agents, threads, models, vector stores, datasets, and evaluations will be unavailable. We will delete all user-specific data pertaining to threads, logins, environment variables, secrets, vector stores, and evaluations shortly after the website shuts down. Most agent code and datasets have always been public on NEAR AI; as such, we will keep paying for the S3 bucket that stores that information for the foreseeable future, and can serve individual requests for that data at our discretion upon request.
What will happen to my NEAR AI agents?
When the NEAR AI Hub goes offline on October 31, 2025, you'll lose access to your agents. Please use the nearai registry download command or the "Export" button to download a copy of your agent code. From there, we recommend adapting it into the AI framework of your choice and hosting it using the infrastructure of your choice. Please reach out on the NEAR AI Telegram group if you'd like migration advice, as we've been working with many users to help migrate their agents to other frameworks.
What will happen to NEAR AI inference/completions API?
The old NEAR AI completions APIs will go offline on October 31, 2025. We recommend NEAR AI Cloud, our new, fully private & confidential AI API, with a similar OpenAI-compatible API. NEAR AI Cloud is in active alpha development and as of October 2025 has not been publicly launched yet; we are disclosing it to the NEAR AI community to maintain service continuity and show you what's coming soon.
If you urgently need production-ready AI services in advance of the NEAR AI Cloud launch, we also like:
- Fireworks AI, which was powering NEAR AI Hub
- RedPill, another confidential AI API
What will happen to NEAR AI Chat and NEAR AI Assistant?
The NEAR AI Assistant at https://chat.near.ai has been redirected to our new private, confidential AI chat interface.
Who can I contact for more information or complaints?
We're available on the NEAR AI Telegram group.
---
The NEAR AI hub & agent framework has been deprecated, and the server components will be shut off on October 31, 2025.





NEAR AI ~is~ was a distributed system for building, deploying, and managing AI agents with the goal of making open source and user-owned AGI.
NEAR AI Components
- [NEAR AI Hub](./hub/README.md): Central hub for model serving, agent registry, and running agents
- TEE Runner: Confidential execution environment for NEAR AI agents and inference
- [AWS Runner](./aws_runner/README.md): Lambda-based execution environment for NEAR AI agents
- Agent System: Build and run AI agents with built-in tools and environment isolation
- Worker System: Distributed job execution and scheduling
- Model Fine-tuning: Support for fine-tuning LLMs
Directory Structure
nearai/
├── aws_runner/ # Lambda-based AI agent execution
│ ├── Dockerfile # Container for running agents
│ └── frameworks/ # Framework-specific requirements
├── hub/ # Central hub for registering and running agents and models
│ ├── alembic/ # Database migrations
│ └── api/ # API endpoints
├── nearai/ # Core library
│ ├── agents/ # Agent system implementation
│ │ ├── agent.py # Base agent class
│ │ └── environment.py # Agent environment
│ ├── cli.py # Command-line interface
│ └── config.py # Configuration management
├── worker/ # Distributed job execution
├── etc/ # Configuration and setup
│ ├── finetune/ # Model fine-tuning configs
│ └── hosts_lambda.txt # Cluster configuration
└── e2e/ # End-to-end tests
Agent Creation Quick Start
Requirements
- Python 3.11 _(3.12+ currently not supported)_
- Git
- Docker (for local agent testing)
Installation
pip
bash
python3 -m pip install nearai
Verify installation:
bash
nearai version
local
bash
git clone [email protected]:nearai/nearai.git && cd nearai && ./install.sh
Or, to install to a virtual environment with poetry:
bash
python3 -m uv sync
uv run nearai version
Or you can use pip:
bash
python3 -m pip install -e .
Verify installation:
bash
nearai version
---
Log In
Login to NEAR AI with your NEAR Account. If you don't have one, we recommend creating a free account with Meteor Wallet.
bash
nearai login
---
Useful Commands
1. Create an agent:
bash
nearai agent create
2. Run agent locally:
bash
nearai agent interactive
3. Deploy to NEAR AI Developer Hub:
bash
nearai registry upload
Documentation
Updating
bash
cd nearai
git pull
python3 -m pip install -e . # If dependencies changed
Contributing
Want to help shape the future of AI? Join our community and contribute! 🚀
- 🐛 Report bugs and suggest features
- 💻 Submit pull requests
- 📖 Improve documentation
- 🤝 Help other users in the community
- 🌟 Star our repository
-
ironclaw
IronClaw is an Agent OS focused on privacy, security and extensibility
Rust ★ 12k 1h agoExplain → -
torchfold
Tools for PyTorch
Python ★ 225 3y agoExplain → -
program_synthesis
Program Synthesis
Jupyter Notebook ★ 150 5y agoExplain → -
nearai ▣
Deprecated – check out our newest work at https://near.ai
Python ★ 115 6mo agoExplain → -
private-ml-sdk
Run LLMs and agents on TEEs leveraging NVIDIA GPU TEE and Intel TDX technologies.
Python ★ 104 8mo agoExplain → -
near-mcp
An MCP server for the NEAR blockchain
TypeScript ★ 31 1y agoExplain → -
aitp
AITP: Agent Interaction & Transaction Protocol
Python ★ 25 1y agoExplain → -
smart-contract-example
Smart Contract with Autonomous AI Execution
Rust ★ 20 1y agoExplain → -
nearai_langchain
NearAI integration for LangChain
Python ★ 12 1y agoExplain → -
official-agents
No description.
Python ★ 12 4mo agoExplain → -
openclaw-nearai-worker
Worker for Hosting OpenClaw in NEAR AI Cloud
Rust ★ 9 6h agoExplain → -
private-chat
Private Chat UI
TypeScript ★ 7 4d agoExplain → -
cloud-api
NEAR AI Cloud API
Rust ★ 6 1d agoExplain → -
benchmarks
Benchmarks for agents
Rust ★ 5 14h agoExplain → -
ironhub
WASM tools and SKILL.md skills for the IronClaw agent runtime.
TypeScript ★ 5 3d agoExplain → -
papers
Proof of Response
★ 4 1y agoExplain → -
near.ai
No description.
JavaScript ★ 4 7y agoExplain → -
inference-proxy
vllm proxy written in Rust
Rust ★ 3 1d agoExplain → -
chat-api
Private Chat API
Rust ★ 3 1d agoExplain → -
near-kms
NEAR KMS for TEE applications
Rust ★ 3 4mo agoExplain → -
compose-manager
No description.
Rust ★ 2 1d agoExplain → -
agents-market-sdk
SDK for hiring AI agents on the NEAR Agent Marketplace. Node.js backend middleware + React components + embeddable widget for any frontend.
JavaScript ★ 2 1mo agoExplain → -
verified-proxy
TEE-verifying forward proxy for NEAR AI inference endpoints
Python ★ 2 3mo agoExplain → -
analytics
No description.
Python ★ 2 10mo agoExplain → -
docs
NEAR AI Cloud Documentation
CSS ★ 1 2d agoExplain → -
ironclaw-desktop-app
Native macOS desktop client for IronClaw — a NEAR AI Cloud agent in a Tauri v2 shell. One-click sign-in, bundled local sidecar, on-device PDF/Office extraction with offline OCR, app connectors, and downloadable generated work product.
JavaScript ★ 1 10m agoExplain → -
ironclaw-dind
No description.
Shell ★ 1 16d agoExplain → -
ironclaw-docs
Documentation for IronClaw
★ 1 2mo agoExplain → -
mintlify-docs
No description.
MDX ★ 1 3mo agoExplain → -
vllm-proxy
No description.
Python ★ 1 4mo agoExplain → -
vllm-router
No description.
Python ★ 1 6mo agoExplain → -
opencode
No description.
Shell ★ 1 7mo agoExplain → -
research-agent-example
No description.
HTML ★ 1 10mo agoExplain → -
oaa-2
Open Agent Alliance Website
HTML ★ 1 1y agoExplain → -
pay-per-prompt ▣
Pay for LLM inference using payment channels
Rust ★ 1 1y agoExplain → -
competition-0.5b-template
No description.
★ 1 1y agoExplain → -
cvm-compose-files
No description.
Ruby ★ 0 1h agoExplain → -
nearai-cloud-verifier
NEAR AI Cloud Verifier
Python ★ 0 2d agoExplain → -
private-chat-ios
NEAR Private Chat native iOS client (SwiftUI, iOS 17+)
Swift ★ 0 4d agoExplain → -
near-ai-web
Near.AI Website & CMS (Blog)
TypeScript ★ 0 5d agoExplain → -
cvm-ingress
No description.
Shell ★ 0 10d agoExplain → -
dstack-vpc ⑂
No description.
Shell ★ 0 15d agoExplain → -
house-of-stake-contracts ⑂
No description.
Rust ★ 0 18d agoExplain → -
genai-benchmark
No description.
Rust ★ 0 1mo agoExplain → -
models.dev ⑂
An open-source database of AI models.
★ 0 1mo agoExplain → -
near-ironclaw
No description.
TypeScript ★ 0 1mo agoExplain → -
ironclaw-e2e-test ⑂
IronClaw is OpenClaw inspired implementation in Rust focused on privacy and security
★ 0 2mo agoExplain → -
dstack-ingress-vpc
No description.
Python ★ 0 1mo agoExplain → -
dstack-vpc-client
No description.
Shell ★ 0 3mo agoExplain → -
dstack-openssh-installer ⑂
No description.
Shell ★ 0 4mo agoExplain → -
certbot-sync ▣
No description.
Shell ★ 0 4mo agoExplain → -
tee-attestation-server ▣
Lightweight TEE attestation sidecar for dstack CVM deployments
Rust ★ 0 4mo agoExplain → -
openclaw ⑂
Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞
TypeScript ★ 0 4mo agoExplain → -
huggingface.js ⑂
Use Hugging Face with JavaScript
★ 0 2mo agoExplain → -
SkyRL ⑂
SkyRL: A Modular Full-stack RL Library for LLMs
Python ★ 0 7mo agoExplain → -
api.near.ai-retirement
No description.
JavaScript ★ 0 7mo agoExplain → -
gorilla ⑂
Gorilla: Training and Evaluating LLMs for Function Calls (Tool Calls)
Python ★ 0 7mo agoExplain → -
simple-evals ⑂
No description.
★ 0 10mo agoExplain → -
lm-evaluation-harness ⑂
A framework for few-shot evaluation of language models.
★ 0 8mo agoExplain → -
NeMo ⑂
A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)
★ 0 2y agoExplain → -
pr-focus
PR Review app for the Age of AI
TypeScript ★ 0 11mo agoExplain → -
oaa ⑂
Open Agent Alliance Website
★ 0 1y agoExplain → -
nearai-auth
No description.
TypeScript ★ 0 1y agoExplain → -
assessment-technical-py-scenario1
No description.
Python ★ 0 1y agoExplain → -
assessment-technical-ts-scenario1
Template repo for technical assessments in TypeScript
TypeScript ★ 0 1y agoExplain → -
payment-channels
Payment channels on NEAR
TypeScript ★ 0 1y agoExplain → -
ui
A React component library that implements the official design system of NEAR AI.
TypeScript ★ 0 1y agoExplain → -
marketing
The landing and marketing pages for NEAR AI (blogs, RFPs, etc)
TypeScript ★ 0 1y agoExplain → -
llm.c ⑂
LLM training in simple, raw C/CUDA
★ 0 1y agoExplain → -
Megatron-LM-for-MegaBlocks ⑂
Ongoing research training transformer models at scale
★ 0 2y agoExplain → -
Megatron-LM ⑂
Ongoing research training transformer models at scale
★ 0 2y agoExplain → -
vllm ⑂
A high-throughput and memory-efficient inference and serving engine for LLMs
Python ★ 0 2y agoExplain →
No repos match these filters.