Getting Started

Contents

• Prerequisites
• Running the Web App
• Installing the CLI
• First Decomposition
  • Web app
  • CLI
• Setting a Default Backend
• What Happens Next

---

Prerequisites

Requirement	Purpose	Minimum Version
.NET SDK	Runs OpenTransmute	10.0
claude CLI	Default AI backend	Latest
Git	Decomposing remote repositories	Any

Verify your setup:

dotnet --version   # should print 10.x.x
claude --version   # should print claude CLI version
git --version

The claude CLI is only required if you intend to use the Claude Agent (ClaudeCode) backend. For OpenAI or Ollama backends, Git and .NET are sufficient.

---

Running the Web App

git clone https://github.com/OpenTransmute/OpenTransmute
cd OpenTransmute/src/OpenTransmute
dotnet run

Open http://localhost:5000 in your browser. The SQLite database is created automatically at DB/opentransmute.db in the current directory on first run.

---

Installing the CLI

The CLI (otx) is a .NET global tool that shares the same database as the web app.

# From the repo root — build and install globally
dotnet tool install --global --add-source ./src/OpenTransmute.Cli otx

Confirm it works:

otx --help

To run without installing:

dotnet run --project src/OpenTransmute.Cli -- --help

Both the web app and the CLI read and write the same DB/opentransmute.db in your working directory. Run them from the same directory to share data.

---

First Decomposition

Web app

1. Open the Decompose screen (home page).
2. Paste a git URL or a local folder path into the Source field.
3. Leave all other settings at their defaults.
4. Click Run.

The job detail page opens immediately and streams live output as each phase runs.

CLI

otx decompose https://github.com/org/repo

The CLI streams phase log output to stdout in real time. When complete, output files are in Output/Decomposition/<project>/.

---

Setting a Default Backend

The default backend is OpenAI. To switch to Claude (recommended) or Ollama, use otx settings:

# Use Claude Code as the default
otx settings --orchestrator ClaudeCode

# Use Ollama with specific models
otx settings \
  --orchestrator Ollama \
  --thick-model llama3.1:70b \
  --regular-model llama3.1:8b \
  --thin-model llama3.2:3b

# Use OpenAI
otx settings \
  --orchestrator OpenAI \
  --thick-model gpt-4o \
  --regular-model gpt-4o-mini \
  --thin-model gpt-4o-mini

See AI Backends for full setup instructions for each backend.

---

What Happens Next

After a decomposition completes:

• Specification files land in Output/Decomposition/<project>/ (eight Markdown files).
• The Phase 6 inventory is automatically parsed into the SQLite database.
• You can browse the inventory in the web app’s Inventory screen, or via otx inventory.
• You can run a Compose or Transmute job to synthesise new designs from the extracted knowledge.

Continue reading:

• Decompose workflow — detailed guide to the eight phases
• Compose workflow — building new designs from inventory items
• Transmute workflow — re-implementing a whole project in a new stack
• CLI Reference — every otx command