How to use OpenCode with Claude, DeepSeek and GPT via RuAPI

OpenCode is an open-source AI coding agent that runs in your terminal: it reads and edits project files, runs commands and sees a task through from start to finish. It isn’t tied to a single model vendor, so it connects to RuAPI directly. RuAPI gives you access to Claude, GPT, Gemini, DeepSeek, GLM, Qwen and Kimi with a single API key, and works from Russia without a VPN. You top up your balance in USDT; no foreign-issued card required.

How it works

In OpenCode, models are connected through providers defined in opencode.json. For RuAPI you need two:

OpenCode provider	Base URL	Models	Why
`anthropic` (built-in)	`https://www.ruapi.ai/v1`	Claude: `claude-opus-5-5`, `claude-sonnet-5-5`, `claude-haiku-5-5`	Claude’s native format: prompt caching works, so re-sent context costs at least 10× less
`ruapi` (custom, OpenAI-compatible)	`https://www.ruapi.ai/v1`	DeepSeek, GPT, Gemini, GLM, Qwen, Kimi and others	One URL for every other model

Before you start

RuAPI API key

Sign up at www.ruapi.ai and top up your balance (how to top up). Then go to Console → API Keys and click Create API key. Keys start with sk-.

OpenCode installed

On macOS and Linux:

curl -fsSL https://opencode.ai/install | bash

Or via npm on any OS, including Windows:

npm install -g opencode-ai

On Windows, the OpenCode team recommends running it under WSL.

No bank card needed: you can top up your balance in USDT. If you don’t have USDT yet, follow our step-by-step guide to buying it with rubles. Foreign-issued cards are accepted too.

Step 1. Store your key in an environment variable

To keep the key out of your config file, OpenCode will read it from the RUAPI_API_KEY environment variable.

macOS / Linux
Windows (PowerShell)

Add the line to ~/.zshrc (or ~/.bashrc if you use bash), with your own key:

echo 'export RUAPI_API_KEY="sk-your-key"' >> ~/.zshrc

Then open a new terminal window for the change to take effect.

setx RUAPI_API_KEY "sk-your-key"

Then open a new PowerShell window for the change to take effect.

Step 2. Define the providers in opencode.json

Open (or create) the global config file ~/.config/opencode/opencode.json and paste:

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "anthropic": {
      "options": {
        "baseURL": "https://www.ruapi.ai/v1",
        "apiKey": "{env:RUAPI_API_KEY}"
      }
    },
    "ruapi": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "RuAPI",
      "options": {
        "baseURL": "https://www.ruapi.ai/v1",
        "apiKey": "{env:RUAPI_API_KEY}"
      },
      "models": {
        "deepseek-flash": { "name": "DeepSeek Flash" },
        "deepseek-v4-pro": { "name": "DeepSeek V4 Pro" },
        "gpt-5.4": { "name": "GPT-5.4" },
        "glm-5.3": { "name": "GLM-5.3" }
      }
    }
  },
  "model": "anthropic/claude-sonnet-5-5"
}

What this does:

anthropic is OpenCode’s built-in Claude provider. The config only points it at RuAPI and passes in your key, so requests use Claude’s native format, with caching.
ruapi is a custom provider for all other models. To add a model, add a line to the models block, e.g. "kimi/kimi-k3": { "name": "Kimi K3" }. You’ll find the exact model names on the pricing page.
model sets the default model in provider/model format.

If your project folder also has an opencode.json, OpenCode merges the two files, and project settings take precedence. Don’t put the key directly in a project file — use {env:RUAPI_API_KEY} as in the example.

Step 3. Start OpenCode and pick a model

Go to your project folder and start the agent:

cd path/to/project

opencode

In OpenCode, type /models and pick a model — for example, anthropic/claude-sonnet-5-5 or ruapi/deepseek-flash. Then give it a task:

Add error handling to the parseConfig function in config.ts

To confirm that requests are going through RuAPI, open Console → Usage Logs: you’ll see an entry for each request with the model and cost. For Claude, from the second or third step onward the entry details show how many tokens were read from the cache.

Which model to pick

Task	Model
Complex tasks, design work	`anthropic/claude-opus-5-5`
Everyday coding	`anthropic/claude-sonnet-5-5`
Lots of routine edits on a tight budget	`ruapi/deepseek-flash` (DeepSeek caches automatically)

Troubleshooting

401 / Invalid key

Make sure RUAPI_API_KEY is available in the terminal where OpenCode runs: after creating the variable, open a new window.
baseURL must be https://www.ruapi.ai/v1 for both providers.

A Claude model is missing from /models

OpenCode takes the model list for the built-in anthropic provider from its own catalog. Update OpenCode; if the model is still missing, add a models section to the anthropic block with the name you need, e.g. "claude-opus-5-5": { "name": "Claude Opus 5.5" }.

Model not found (404 or 503)

The model name must match the one on RuAPI exactly — copy it from the pricing page. Pick Claude models from the anthropic provider and all other models from ruapi.

Config file error

Make sure the JSON is valid: all quotes are double quotes, and there’s no trailing comma after the last item in a block.

400: The reasoning_content in the thinking mode must be passed back

This can happen with DeepSeek in reasoning mode if you switch models mid-task. In most cases RuAPI automatically fixes and retries the request. If the error persists, start a new session.

What’s next

Claude Code and Codex — other terminal agents.
Switching from Cursor — a comparison of coding agents and step-by-step migration.
Models and prices — the full list of available models.

Getting Started

Guides

API Reference

Model guides

Integrations

Capabilities

Recipes

Legal

How to use OpenCode with Claude, DeepSeek and GPT via RuAPI

How it works

Before you start

Step 1. Store your key in an environment variable

Step 2. Define the providers in opencode.json

Step 3. Start OpenCode and pick a model

Which model to pick

Troubleshooting

What’s next

​How it works

​Before you start

​Step 1. Store your key in an environment variable

​Step 2. Define the providers in opencode.json

​Step 3. Start OpenCode and pick a model

​Which model to pick

​Troubleshooting

​What’s next

How it works

Before you start

Step 1. Store your key in an environment variable

Step 2. Define the providers in opencode.json

Step 3. Start OpenCode and pick a model

Which model to pick

Troubleshooting

What’s next