NanoBridge

OpenAI-compatible retry and fallback gateway for Claude Code and AI agent workloads.

Built for developers dealing with:

• unstable AI providers
• model timeouts
• provider failover
• coding workload spikes
• token cost optimization
• multi-provider routing

---

Features

• OpenAI-compatible API endpoint
• Automatic retry handling
• Smart fallback routing
• Multi-provider failover
• Claude Code friendly
• Lightweight proxy architecture
• Fast integration with existing tools

Use Cases

• Reduce downtime during provider overload
• Route requests across multiple AI APIs
• Improve coding workflow stability
• Lower AI coding costs with smart routing
• Add fallback support for AI agents

Works With

• Claude Code
• OpenAI SDK
• Cursor
• Cline
• Roo Code
• Continue
• AI agent frameworks
• MCP workflows

⚡ What this solves

• Cursor setup fails with DeepSeek
• OpenAI-compatible API confusion
• Wrong base URL / model issues
• Slow or broken integrations

This repo gives you a clean working setup.

---

🚀 Quick Start

1. Cursor Config

{
  "apiKey": "YOUR_API_KEY",
  "baseURL": "https://platform.nanobridge.net/v1",
  "model": "deepseek-v4-pro"
}
---

# Features

- OpenAI-compatible API
- DeepSeek V4 Pro support
- Streaming support
- Long-context reasoning
- Coding-oriented inference routing
- Retry optimization
- Crypto payment support
- AI agent compatibility

---

# Supported Models

Currently available:

- deepseek-v4-pro
- deepseek-v3
- deepseek-reasoner

More models coming soon.

---

# Supported Tools

Works with:

- Cursor
- Cline
- Continue
- Aider
- OpenAI SDK
- LangChain
- LiteLLM
- AI agents

---

# API Endpoint

Base URL:

``` id="j1k2vs"
https://platform.nanobridge.net/v1

API Docs:

https://platform.nanobridge.net/api-docs

---

Quick Start

Python

Install SDK:

pip install openai

Example:

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://platform.nanobridge.net/v1"
)

response = client.chat.completions.create(
    model="deepseek-v4-pro",
    messages=[
        {
            "role": "user",
            "content": "Write a FastAPI middleware example"
        }
    ],
    stream=False
)

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

---

cURL Example

curl https://platform.nanobridge.net/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "deepseek-v4-pro",
    "messages": [
      {
        "role": "user",
        "content": "hello"
      }
    ]
  }'

---

Cursor Setup

Step 1 — Open Cursor Settings

Go to:

Cursor → Settings → Models

Enable:

OpenAI Compatible API

---

Step 2 — Configure API

Use:

{
  "apiKey": "YOUR_API_KEY",
  "baseURL": "https://platform.nanobridge.net/v1",
  "model": "deepseek-v4-pro"
}

---

Recommended AI Coding Workflow

DeepSeek V4 Pro works especially well for:

• refactoring
• bug fixing
• code explanation
• multi-file edits
• long-context coding
• reasoning-heavy workflows
• autonomous coding agents

---

Streaming Support

Streaming is fully supported.

Recommended for:

• lower perceived latency
• coding copilots
• interactive workflows
• long generations

---

OpenAI Compatibility

Nanobridge is designed to be drop-in compatible with OpenAI-style APIs.

Supported:

• /chat/completions
• streaming
• system prompts
• multi-turn conversations

This makes integration easy with most modern AI developer tools.

---

Why Not Just Use OpenRouter?

Some developers prefer:

• lower latency
• coding-focused routing
• simplified infrastructure
• stable long-context workflows
• direct DeepSeek optimization
• crypto payment support

Nanobridge focuses specifically on AI coding workloads and developer workflows.

---

AI Agent Optimization

Optimized for:

• coding agents
• autonomous workflows
• multi-step reasoning
• tool-calling systems
• development automation

---

Example Prompt Ideas

Try inside Cursor:

Refactor this codebase

Find concurrency issues

Optimize this SQL query

Explain this architecture

---

Best Practices

For best coding performance:

• use focused prompts
• split large tasks
• restart long sessions periodically
• prefer streaming mode
• use explicit coding instructions

---

Common Issues

Timeout Problems

Possible causes:

• overloaded providers
• unstable routing
• long reasoning chains

Possible solutions:

• retry requests
• shorten prompts
• restart conversation sessions

---

Invalid Base URL

Make sure the Base URL is:

https://platform.nanobridge.net/v1

NOT:

https://platform.nanobridge.net/api-docs

---

Performance Goals

Current optimization targets:

• lower latency
• stable streaming
• reliable coding workflows
• reduced inference failures

---

Current Status

Project status:

• Active development
• Early production usage
• API stability improvements ongoing

---

Roadmap

Planned improvements:

• smarter routing
• fallback optimization
• additional models
• lower latency
• enhanced AI agent support
• coding workflow tuning

---

Community

Website: https://www.nanobridge.ai

Platform: https://platform.nanobridge.net

API Docs: https://platform.nanobridge.net/api-docs

GitHub: https://github.com/nanobridgerafa

---

Disclaimer

Nanobridge is an independent inference gateway project and is not affiliated with OpenAI or DeepSeek.