Documentation - Gwarden AI

Agents

Choose your tool

Get a key

Open the bot and press «Get API» - the key looks like GWAR-XXXX. Your free 7-day trial starts on signup (balance refills every 5h).

Cabinet Bot

2

CLI coding agents

Works with every major CLI agent. Install for your OS, then pick your tool. All of them speak the OpenAI- or Anthropic-compatible protocol - point them at https://gwarden.su.

# macOS
brew install node
# Linux (Debian / Ubuntu)
sudo apt install -y nodejs npm
# Windows (PowerShell, admin)
winget install OpenJS.NodeJS.LTS

# all OS: Claude Code + Codex + opencode
npm install -g @anthropic-ai/claude-code @openai/codex opencode-ai

# aider (Python)
python -m pip install aider-install && aider-install

Replace GWAR-XXXX with your key and MODEL with any model id - the full list is via /v1/models and in the table on the main page. Examples: claude-fable-5-1, gpt-5.6-sol, grok-4.6, kimi-k3, glm-5.3. Tools / agentic loops are fully supported.

cli

Claude Code

One command - works out of the box

# macOS / Linux
export ANTHROPIC_BASE_URL="https://gwarden.su"
export ANTHROPIC_AUTH_TOKEN="GWAR-XXXX"
export ANTHROPIC_MODEL="MODEL"
claude

# Windows PowerShell
$env:ANTHROPIC_BASE_URL = "https://gwarden.su"
$env:ANTHROPIC_AUTH_TOKEN = "GWAR-XXXX"
$env:ANTHROPIC_MODEL = "MODEL"
claude

cli

Codex CLI (OpenAI)

~/.codex/config.toml
(Windows: C:/Users/you/.codex/config.toml)

model = "MODEL"
model_provider = "gwarden"

[model_providers.gwarden]
name = "Gwarden AI"
base_url = "https://gwarden.su/v1"
env_key = "GWARDEN_API_KEY"
wire_api = "responses"

export GWARDEN_API_KEY="GWAR-XXXX"   # Windows: setx GWARDEN_API_KEY "GWAR-XXXX"
codex

cli

opencode

~/.config/opencode/opencode.json

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "gwarden": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Gwarden AI",
      "options": { "baseURL": "https://gwarden.su/v1", "apiKey": "GWAR-XXXX" },
      "models": { "MODEL": { "name": "Gwarden AI" } }
    }
  }
}

opencode --model gwarden/MODEL

cli

aider

# macOS / Linux
export OPENAI_API_BASE="https://gwarden.su/v1"
export OPENAI_API_KEY="GWAR-XXXX"
aider --model openai/MODEL

# Windows PowerShell
$env:OPENAI_API_BASE = "https://gwarden.su/v1"
$env:OPENAI_API_KEY = "GWAR-XXXX"
aider --model openai/MODEL

3

Chat Completions

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "MODEL",
    "messages": [{"role": "user", "content": "Hello"}],
    "stream": false}'

Models list

curl https://gwarden.su/v1/models -H "Authorization: Bearer KEY"

Streaming

Set stream: true. Chunks carry reasoning_content (thoughts) and content (answer) separately.

Old requests with thinking: {{"type": "disabled"}} are rewritten to {{"type": "enabled"}} + reasoning_effort: "low".

4

Claude Code / Anthropic compat

The same API is exposed in the Anthropic Messages format, so Claude Code, Cursor and Cline work out of the box. Point the base URL here and use your key.

curl https://gwarden.su/v1/messages \
  -H "x-api-key: GWAR-XXXX" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "MODEL",
    "max_tokens": 4096,
    "messages": [{"role": "user", "content": "Hello"}]
  }'

5

Subscriptions

Free

Up to $6$ credits

Refills every 5 h to $6$

7-day trial

Tester

$4

Up to $5$ credits

Refills every 5 h to $5$

Term