Your work. Your space.

Find the right Gonu tool for the next step.

Features & demosExplore Engineering Brain workflows and sample screensExploreWeb ChatFor quick questions, planning and everyday workTry nowDesktop WorkerFor local repository work with a dedicated desktop appDownloadVS Code ExtensionFor developers who already work in VS CodeSource betaGonu CLIFor terminal workflows and structured task inspectionSource betaGonu IDEFor developers exploring an integrated Gonu editorNot releasedDocs & demosPractical guides for every Gonu toolLearn
Skip to content
gonu.
Checking account…Try on web

SOFTWARE INTELLIGENCE / WORKFLOW GUIDE

Understand the change.
Keep the reasoning.

Explore 20 focused workflows for your engineering decisions, shared knowledge and controlled experiments.

Open workspace
Illustrative examples · no live task or model usageAll Gonu features →

BOUNDED V1 WORKFLOW

Decision Compiler

Turn a reviewed decision and linked files into workstreams and a controlled engineering plan.

Sources and affected paths are required for specific impact; a plan must be reviewed before execution.

01 / PROVIDE RELEVANT INPUTS

Keep session expiry in the auth service. Select auth.py and a reviewed decision.

02 / INSPECT THE OUTPUT · SAMPLE CONTENT

Affected consumer
main.py imports auth.py
Proposed workstream
Update expiry handling and regression checks

03 / CHOOSE THE NEXT ACTION

Review acceptance criteria, then prepare an engineering plan.

Try with your project Repository setup guide →

What you need

  • Your existing Gonu account and an authorized workspace project.
  • A managed repository snapshot for code and history workflows.
  • Reviewed decisions, consented expertise or real observations for knowledge workflows.
  • An administrator to accept shared policies and merge knowledge.

These workflows share Gonu's identity and engineering runtime. Deterministic analysis does not call an LLM. Agent execution needs configured models and available allowance; offline experiments need the configured sandbox. No model can guarantee causal certainty or error-free code.