Skip to content
FullsendAI
Log inGet started

Agents

Works with
your agent.

Claude Code, Codex, Cursor, Cline and any tool that lets you set a base URL. One command connects them all with every model we serve, Nef and Sifu included, so you can switch models from your tool's picker.

Connect in one commandGet API Key

Connect them all at once.

Run this in a terminal. It finds your coding tools, signs you in through the browser, gives each its own key and sets it up, with a backup of every file it changes.

npx fullsendai connect

Claude Code, Codex, OpenCode, Kilo Code, Cline, Aider and Continue. macOS, Linux and Windows. Undo it all with npx fullsendai disconnect. How it works

Or pick your tool.

To set one up by hand: each opens its setup below. Every tool takes one of two addresses, and both take the same key.

OpenAI-compatible · Cursor, Cline, Aider, Codex CLI and the OpenAI SDK
https://fullsendai.com/v1
Anthropic-compatible · Claude Code and the Anthropic SDKs
https://fullsendai.com

Plus your fs_live_… key and a model ID from the model list.

One-minute setup.

Copy the settings, swap in your fs_live_… key, and run.

Claude Code

Anthropic API
  1. Set the base URL, your key and the models Claude Code should use.
  2. Run claude as usual. Every request goes to FullsendAI.
export ANTHROPIC_BASE_URL=https://fullsendai.com
export ANTHROPIC_AUTH_TOKEN=fs_live_...
export ANTHROPIC_MODEL=nef-4.0
export ANTHROPIC_DEFAULT_HAIKU_MODEL=nef-4.0
claude

Faster: npx fullsendai connect sets up Claude Code for you, with its /model picker.

Full Claude Code guide

Cursor

OpenAI-compatible API
  1. Settings → Models → API Keys: paste your key as the OpenAI API Key.
  2. Turn on Override OpenAI Base URL and enter the base URL.
  3. Add a custom model with a FullsendAI model ID.
Cursor settings
OpenAI API Key      fs_live_...
Override Base URL   https://fullsendai.com/v1
Custom model        nef-4.0

Faster: npx fullsendai connect copies your key and shows exactly these values to paste.

Full Cursor guide

Cline and Roo Code

OpenAI-compatible API
  1. Open the extension’s settings.
  2. Set API Provider to OpenAI Compatible.
  3. Enter the base URL, your key and a model ID.
Extension settings
API Provider   OpenAI Compatible
Base URL       https://fullsendai.com/v1
API Key        fs_live_...
Model ID       nef-4.0

Faster: npx fullsendai connect sets up Cline for you, its VS Code extension too. For Roo Code it shows these values.

Full guide for these

Continue

OpenAI-compatible API
  1. Add FullsendAI as a model in your Continue config.
  2. Pick it in the Continue panel.
~/.continue/config.yaml
models:
  - name: FullsendAI
    provider: openai
    model: nef-4.0
    apiBase: https://fullsendai.com/v1
    apiKey: fs_live_...
    roles: [chat, edit, apply]

Faster: npx fullsendai connect adds every FullsendAI model to Continue for you.

Full Continue guide

Aider

OpenAI-compatible API
  1. Point Aider’s OpenAI settings at FullsendAI.
  2. Start it with a FullsendAI model.
shell
export OPENAI_API_BASE=https://fullsendai.com/v1
export OPENAI_API_KEY=fs_live_...
aider --model openai/nef-4.0

Faster: npx fullsendai connect sets up Aider for you, with every model’s limits and prices.

Full Aider guide

Codex CLI

OpenAI-compatible API
  1. Add FullsendAI as a model provider in ~/.codex/config.toml.
  2. Export your key as FULLSEND_API_KEY and run codex.
model = "nef-4.0"
model_provider = "fullsend"

[model_providers.fullsend]
name = "FullsendAI"
base_url = "https://fullsendai.com/v1"
env_key = "FULLSEND_API_KEY"
wire_api = "responses"

Faster: npx fullsendai connect sets up Codex for you, with FullsendAI’s models in its model menu.

Full Codex CLI guide

OpenCode and Kilo Code

OpenAI-compatible API
  1. Add FullsendAI as a provider in opencode.json, or in ~/.config/kilo/kilo.json for Kilo Code and its extension.
  2. Export your key, run it, and pick a FullsendAI model with /models.
{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "fullsend": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "FullsendAI",
      "options": {
        "baseURL": "https://fullsendai.com/v1",
        "apiKey": "{env:FULLSEND_API_KEY}"
      },
      "models": {
        "nef-4.0": { "name": "Nef 4.0", "tool_call": true }
      }
    }
  },
  "model": "fullsend/nef-4.0"
}

Faster: npx fullsendai connect sets up OpenCode and Kilo Code for you, with every FullsendAI model.

Full guide for these

OpenAI SDK

OpenAI-compatible API
  1. Pass the base URL and your key to the client.
  2. Call Chat Completions with any model ID.
from openai import OpenAI

client = OpenAI(
    base_url="https://fullsendai.com/v1",
    api_key="fs_live_...",
)
reply = client.chat.completions.create(
    model="nef-4.0",
    messages=[{"role": "user", "content": "Write a debounce"}],
)
Full OpenAI SDK guide

Vercel AI SDK

OpenAI-compatible API
  1. Create an OpenAI-compatible provider for FullsendAI.
  2. Use it like any other model.
TypeScript
import { createOpenAICompatible } from "@ai-sdk/openai-compatible";
import { generateText } from "ai";

const fullsend = createOpenAICompatible({
  name: "fullsend",
  baseURL: "https://fullsendai.com/v1",
  apiKey: process.env.FULLSEND_API_KEY, // fs_live_...
});

const { text } = await generateText({
  model: fullsend("nef-4.0"),
  prompt: "Write a SQL migration",
});
Full Vercel AI SDK guide

LangChain

OpenAI-compatible API
  1. Use ChatOpenAI with the FullsendAI base URL.
  2. Everything else in your chain stays the same.
Python
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    model="nef-4.0",
    base_url="https://fullsendai.com/v1",
    api_key="fs_live_...",
)
Full LangChain guide

Setup questions.

The short answers. The docs have the long ones.

Yes. npx fullsendai connect puts every FullsendAI model in each tool's model picker, Nef and Sifu included, and npx fullsendai sync adds new ones as they go live. Pick a different model whenever you like, or run npx fullsendai model <id> to switch every tool at once. It's one key and one balance either way.

Anything that lets you set a base URL and an API key: Claude Code, Codex, OpenCode, Cursor, Cline, Roo Code, Kilo Code, Continue, Aider, the OpenAI and Anthropic SDKs, LangChain, the Vercel AI SDK and more.

Yes. Paste your FullsendAI key into the tool you already use and it works unchanged, streaming, tool calls and images included. npx fullsendai connect sets most of them up in one go; the Connect agents page shows each tool.

Yes. FullsendAI serves the Responses API alongside Chat Completions and Anthropic Messages, streamed or not, with tool calls. Codex uses it, and so can anything built on OpenAI’s Agents SDK.

No. Cursor always runs Tab on its own models. Chat and agent requests go to FullsendAI once you set the base URL and key.

Your agent, better models.

Sign up in a minute and pay only for what you use.

Get API Key