Skip to content

opencode.md

Latest commit

 

History

History
151 lines (105 loc) · 4.08 KB

opencode.md

File metadata and controls

151 lines (105 loc) · 4.08 KB
Raw

OpenCode + Eden AI

Connect OpenCode to Eden AI and route your terminal agent through 500+ models with smart routing, automatic fallbacks, and a single API key.

Get your Eden AI key at app.edenai.run → API Keys.

Follow this tuto in the same time with our video tuto on Youtube

1. System dependencies

On a fresh Linux machine (Ubuntu / Debian / WSL):

sudo apt update
sudo apt install -y curl git python3-venv

2. Install Node.js

OpenCode requires Node.js 22.14+ and npm 9.8+.

sudo apt install -y nodejs npm

# Verify
node --version   # v22.14.x or newer
npm  --version   # 9.8.x or newer

If the packaged version is too old, install via nvm or nodesource.

3. Install OpenCode

Pick one of the two methods below.

# Option A — official install script
curl -fsSL https://opencode.ai/install | bash

# Option B — via npm
npm install -g opencode-ai

Reload your shell so the opencode binary is on your PATH:

source ~/.bashrc

4. Add Eden AI as a custom provider

OpenCode stores its config at ~/.config/opencode/opencode.json. Open it and add the Eden AI provider with the @edenai smart-router model:

nano ~/.config/opencode/opencode.json
{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "edenai": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Eden AI",
      "options": {
        "baseURL": "https://api.edenai.run/v3",
        "apiKey": "YOUR_EDEN_AI_API_KEY"
      },
      "models": {
        "@edenai": {
          "name": "@edenai (Smart Router)"
        }
      }
    }
  }
}

Why @edenai?

@edenai is Eden AI's smart router. For every request, it automatically picks the best-performing model based on quality, latency, cost and live provider health — so your OpenCode agent always uses the right model without a config change.

How smart routing works →

5. Start OpenCode and pick the model

Launch the TUI:

opencode

Then select the Eden AI provider and model:

  • Run /models and pick @edenai (Smart Router) under Eden AI.
  • If you'd rather authenticate interactively instead of pasting the key into the config, run /connect, scroll to Other, enter edenai as the provider ID, and paste your Eden AI key when prompted.

That's it. Every LLM call now flows through Eden AI — with built-in fallbacks, unified billing, and per-request cost visibility.

Using specific models

@edenai handles routing automatically, but you can target a specific model by adding it as an extra entry under models, keyed by provider/model:

"models": {
  "@edenai": { "name": "@edenai (Smart Router)" },
  "anthropic/claude-sonnet-4-5": { "name": "Claude Sonnet 4.5" }
}

Common choices:

Model Best for
@edenai Default — smart routing across all providers
anthropic/claude-sonnet-4-5 Balanced reasoning
anthropic/claude-opus-4-7 Maximum capability
anthropic/claude-haiku-4-5 Fast and cheap
openai/gpt-5 OpenAI's frontier model
google/gemini-2.5-pro Long context, multimodal

Troubleshooting

401 Unauthorized

Verify your Eden AI key — grab a fresh one at app.edenai.run → API Keys.

model not found

Ensure the model key uses the provider/model format (or @edenai for the smart router). No bare model names.

Connection issues

Confirm the base URL is exactly https://api.edenai.run/v3 and check Eden AI status at app-edenai.instatus.com.

Uninstall

npm uninstall -g opencode-ai
rm -rf ~/.config/opencode

Next steps