PRACTICAL ANSWERS

What do you want to know?

Start with your agent and a small task. Here’s what comes with it, what you can change, and when Cloud helps.

Getting started

Is Weft another coding agent?

You keep talking to your agent. It uses Weft to organize tasks, run workflows, and keep project records. Use the CLI directly whenever useful.

Is OSS free? Do I need an account?

OSS is free and open source under Apache 2.0. Local tasks, memory, workers, workflows, and checks need no Weft account or Cloud subscription. Your agent’s model access is separate.

How will I get started?

At public launch, install Weft, connect your coding agent, and initialize a project. Bring a small change and checks you already use. Public installation and repository access will be announced here.

Can I use it through Dots or another assistant?

Your assistant can delegate to a coding agent that uses Weft. Direct access needs the ability to run the CLI or connect to Weft’s MCP tools. Guided setup supports Claude Code, Codex, and Cursor; direct Dots integration has not been tested.

Agents, memory, and your process

Does it have a memory?

Yes. The local knowledge base keeps decisions and their reasons with your project. Ask your agent to search it before a significant change, ask you when an answer is missing, and record the result for later sessions.

Can I run multiple agents?

Yes. Managed local workers can run independent tasks concurrently in separate worktrees. Each task follows its configured workflow and required checks.

Can I choose different models?

Yes. You or your agent choose the agents and models for the work. Use a lower-cost model for a bounded change and a stronger model for harder work or supervision. Keep your acceptance criteria and checks.

Can I change the workflow?

Yes. Ask your coding agent to write or edit a YAML workflow: change stages, their order, and where checks must pass. Start with the included development flow. Build your own plugin when you need new tools or executable checks.

I already have a software factory. Why use Weft?

Keep your agents, skills, and repository checks. Try one task through Weft. Reuse its workflow engine, work records, and managed workers instead of maintaining that machinery yourself. Your custom integrations and checks still need care.

Why save a workflow if my agent can figure it out?

Your agent still solves the task. The saved process gives it stages and checks to reuse, while project memory carries prior decisions. Inspect the task history and adjust the process as you learn.

Checks and optional Cloud

What does Cloud add?

Hosted project records shared across people and machines, plus AI review of plans, designs, and code. Add it for shared work, for review while working alone, or both. The local OSS product works independently.

Does a passing check prove the software is correct?

A check establishes what it tests. OSS runs your configured build, tests, and structural checks. Choose meaningful acceptance criteria and inspect the results. AI review can add another perspective.

TRY ONE TASK

A small change is enough to start.

Use the included workflow and your existing checks. Adjust the process after you’ve seen it run.

Follow a first task →
Setup and launch availability

Once public installation is available, install the Weft binary and register it with your MCP client. Guided setup supports Claude Code, Cursor, and Codex. Other agents can use Weft through their MCP integration or the CLI.

# After public launch and installation:
weft install --client claude-code
# or:
weft install --client codex
weft init

Restart your assistant, ask it to create a Weft project and a small task, and start with checks you already trust. Your agent remains the interface; you can use the CLI directly whenever useful.

OSS launch is being prepared. Public installation and repository access will be announced here. You can explore the process now and ask for launch updates.