MegaBrain University · guide

OpenWork Custom LLM Setup: the Open-Source Claude Cowork Alternative on Any Model

Updated November 11, 2026 · 10 min setup

What is OpenWork

OpenWork is an open-source desktop app for macOS, Windows and Linux, built as an alternative to Claude Cowork. It gives non-developers an agent that works with files and folders, documents, spreadsheets and the web, with shared skills and MCP connections for a team. Under the hood it runs the OpenCode engine.

Because the engine is OpenCode, OpenWork reads OpenCode’s provider config. Any OpenAI-compatible API can be added as a custom LLM, so a team can run OpenWork on MegaBrain with one balance and switch between Claude, GPT and cheaper open models per task.

Setup, step by step

1. Step 1. Get a MegaBrain API key

   Sign up with GitHub, Google or email, add funds on the Balance page and create a key on the API key page. The key works for every text, image and video model, and each request is paid from your balance — no subscription.

   Put the key in an environment variable so it never lands in a config file you share:

   # ~/.zshrc or ~/.bashrc
   export MEGABRAIN_API_KEY="your-megabrain-key"

   Copy

   curl https://getmegabrain.com/api/gateway/v1/chat/completions \
     -H "Authorization: Bearer $MEGABRAIN_API_KEY" \
     -H "Content-Type: application/json" \
     -d '{"model": "anthropic/claude-sonnet-5.5", "messages": [{"role": "user", "content": "Say OK"}]}'

   Copy

   Check the key before you touch the agent. A JSON reply with choices means the key, the balance and the model all work:

2. Step 2. Install OpenWork

   Download the app for your OS from openworklabs.com/download or the GitHub releases page, install it and open a workspace (a folder the agent will work in).

3. Step 3. Add MegaBrain as a custom LLM

   Create or open opencode.jsonc in the workspace folder (OpenWork’s docs recommend the workspace file over the global one) and add the provider:

   {
     "$schema": "https://opencode.ai/config.json",
     "provider": {
       "megabrain": {
         "npm": "@ai-sdk/openai-compatible",
         "name": "MegaBrain",
         "options": {
           "baseURL": "https://getmegabrain.com/api/gateway/v1",
           "apiKey": "{env:MEGABRAIN_API_KEY}"
         },
         "models": {
           "anthropic/claude-sonnet-5.5": {
             "name": "Claude Sonnet 5.5",
             "limit": { "context": 1000000, "output": 32000 }
           },
           "openai/gpt-6.1-sol": {
             "name": "GPT-6.1 Sol",
             "limit": { "context": 1050000, "output": 32000 }
           },
           "google/gemini-3.8-flash": {
             "name": "Gemini 3.8 Flash",
             "limit": { "context": 1048576, "output": 32000 }
           }
         }
       }
     }
   }

   Copy

4. Step 4. Add the key and reload

   In OpenWork open Settings → Environment → Add variable, name it MEGABRAIN_API_KEY and paste the key (or export it in your shell before starting the app). Then run Reload OpenCode config from the command palette, or restart OpenWork.

5. Step 5. Pick the model and run a task

   Open the model picker in the chat, choose MegaBrain → Claude Sonnet 5.5 and give it a task on a file in the workspace, for example “Summarize every PDF in /contracts into one table”.

Which model to choose

OpenWork needs an endpoint that speaks Chat Completions and a model that handles tool definitions; every model below does. Claude Sonnet 5.5 is the best all-rounder for documents; Gemini Flash is a cheaper choice for long spreadsheets and bulk summaries.

Every model with its price and context window is in the catalog. They all work with the same key.

What you can do

Business use cases

Read every file in ./contracts. Build contracts-summary.xlsx with columns: file, counterparty, signing date, end date, total amount, payment terms, auto-renewal (yes/no), unusual clauses. Flag anything that looks risky in a separate column.

Troubleshooting

FAQ