NanoBridge — Cursor & DeepSeek Integration Guide

OpenAI-compatible retry and failover gateway for Cursor, Claude Code, and AI coding agents.

Built for developers dealing with unstable upstream providers, timeouts, multi-provider routing, coding workload spikes, and cost optimization.

---

What this repo solves

• Wrong Base URL or model name when configuring DeepSeek in Cursor
• Using the console host platform.nanobridge.net as the API endpoint
• Confusing OpenAI-compatible APIs with Anthropic Messages endpoints
• Not knowing which model or regional node to pick

---

Quick Start (Cursor)

1. Get an API key

Sign up and create an API key in the Nanobridge console.

2. Pick a regional Base URL

API requests go to the gateway hosts (not the console):

Region	OpenAI Base (use in Cursor, includes /v1)	Anthropic Base (Cline / Claude Code, etc.)
Germany (default)	https://api.nanobridge.net/v1	https://api.nanobridge.net/anthropic
Singapore	https://api-sg.nanobridge.net/v1	https://api-sg.nanobridge.net/anthropic
United States	https://api-us.nanobridge.net/v1	https://api-us.nanobridge.net/anthropic

The same API key works across regions. Pick the node closest to you for lower latency.

3. Cursor settings

1. Open Cursor → Settings → Models
2. Enable Override OpenAI Base URL (or a similar “OpenAI Compatible API” option)
3. Use:

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

The recommended default model is deepseek-v4-flash (balanced cost and latency). Use deepseek-v4-pro for stronger reasoning.

Note: Some Cursor features may still use built-in channels. The OpenAI override mainly applies to Chat / Composer and other compatible paths—verify with a real request.

Generate config with: python generate_config.py

---

Supported models

Use these IDs in the request model field (matches the console gateway catalog):

Model ID	Notes
deepseek-v4-flash	Recommended default
deepseek-v4-pro	Stronger reasoning
deepseek-v3.2	DeepSeek V3.2
glm-5.1	Alias: glm-5-1
minimax-m2.7	Alias: minimax-m-2-7

---

API endpoints

• OpenAI Chat Completions: POST /v1/chat/completions (or POST /chat/completions)
• Model list: GET /v1/models
• Anthropic Messages: POST /v1/messages (Anthropic base from the table above)
• Streaming: set stream: true in the request body; response is SSE
• Auth: Authorization: Bearer <API_KEY>

Full parameters and examples: chat-completions.md

Official API docs: platform.nanobridge.net/#/api-docs

---

Python (OpenAI SDK)

pip install openai

from openai import OpenAI

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

response = client.chat.completions.create(
    model="deepseek-v4-flash",
    messages=[{"role": "user", "content": "Write a FastAPI middleware example"}],
)
print(response.choices[0].message.content)

For Singapore, change base_url to https://api-sg.nanobridge.net/v1.

---

cURL

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

More examples: api-key-example.sh

---

LiteLLM

export NANOBRIDGE_API_KEY="your-api-key"
# optional: export NANOBRIDGE_API_BASE="https://api-sg.nanobridge.net/v1"

import litellm

response = litellm.completion(
    model="nanobridge/deepseek-v4-flash",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

---

Compatible tools

Tool	Protocol	Configuration
Cursor	OpenAI	Base = https://api*.nanobridge.net/v1
Claude Code	Anthropic	ANTHROPIC_BASE_URL = Anthropic base
Cline / Roo Code / Kilo	Anthropic	Provider = Anthropic, model deepseek-v4-flash
Continue / Aider	OpenAI	Same OpenAI base as above
OpenAI SDK / LangChain	OpenAI	base_url + api_key

Portal integrations: platform.nanobridge.net → Integrations.

---

FAQ

Timeouts

Possible causes: upstream overload, long reasoning chains, network jitter. Shorten prompts, enable streaming, switch to a closer region, or retry.

Wrong Base URL

Correct (gateway)	Wrong
https://api.nanobridge.net/v1	https://platform.nanobridge.net/v1
	https://platform.nanobridge.net/api-docs

platform.nanobridge.net is the console and documentation—it does not serve Chat Completions inference.

Model not found

Do not use retired or unavailable names (e.g. deepseek-v3, deepseek-reasoner). Use the model list in the console.

---

Repository files

File	Description
cursor-deepseek-guide.md	Cursor + DeepSeek focused guide
generate_config.py	Interactive Cursor config generator
chat-completions.md	Chat / Messages API reference
api-key-example.sh	cURL examples
nanobridge-overview.md	Product overview and pricing

---

Links

• Website: https://www.nanobridge.ai
• Console: https://platform.nanobridge.net
• API docs: https://platform.nanobridge.net/#/api-docs
• GitHub: https://github.com/nanobridgerafa

---

Disclaimer

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