Skip to content

Repository files navigation

Ducelis Open

Local-first AI rehearsal for high-stakes conversations.

Ducelis Open is the public reference edition of Ducelis. It is designed to help people rehearse difficult conversations in a structured, scenario-based format rather than use a generic chatbot.

What it is

Ducelis Open is a local-first rehearsal product for scenario-driven dialogue practice, reflection, and reporting.

The initial proof lane focuses on high-stakes education conversations such as:

  • teacher ↔ parent boundary pressure
  • teacher ↔ student disruption and de-escalation
  • teacher ↔ parent fairness or grading challenges

The product is intended to make three things clear:

  1. the scenario has real stakes
  2. the conversation is guided by a structured roleplay contract
  3. the session produces useful reflection value after the interaction

Core experience

A typical session is designed to follow this shape:

  1. choose a scenario
  2. review the situation, goals, and watch-outs
  3. rehearse the conversation with an AI role
  4. complete the session
  5. review summary, evidence, and next-practice guidance

Product principles

  • Local-first - the default product posture is local use and user-controlled session data
  • Model-agnostic - the architecture should not be permanently bound to one model family
  • Reference-quality - the repo should stay understandable, public-safe, and reproducible
  • Rehearsal-first - the product should feel like purposeful practice, not generic chat
  • Evidence-first - summary and report value matter more than decorative surface area

Model support direction

Ducelis Open is initially validated with Gemma 4, but the product is not intended to be hard-wired to a Gemma-only future.

The architecture should support configurable local runtimes and compatible model families over time through stable runtime and model adapter boundaries.

Data direction

Ducelis Open is designed around local-first session ownership.

That means:

  • sessions should remain under user control by default
  • the public repository must not contain real user transcripts or raw private datasets
  • future versions may support user-controlled export paths for evaluation or training workflows
  • this repository should not imply default cloud upload or centralized data collection

Current status

This repository is in the early foundation stage.

The current priority is to define a clean public-safe architecture and then build the smallest complete rehearsal loop with:

  • a stable runtime boundary
  • a text-first conversation flow
  • local session persistence
  • summary / evidence / report basics
  • public-safe demo scenarios

This README does not claim that all of those pieces are already implemented.

Intended project structure

The planned structure is expected to evolve around the following layers:

/apps/open-web
/packages/open-core
/packages/open-runtime
/packages/open-content
/packages/open-reporting
/packages/open-data
/docs

Public repo rules in plain language

This repository should only contain Ducelis Open public-safe materials.

It should not contain:

  • private conversations
  • customer data
  • confidential business planning
  • non-Open Ducelis strategy
  • real user transcripts
  • raw local exports or datasets unless they are clearly synthetic and public-safe

Documentation map

  • AGENTS.md - repository rules for AI coding agents
  • docs/product-scope.md - current v1 scope and non-goals
  • docs/public-repo-guardrails.md - public-safe writing and repo boundaries
  • docs/architecture-overview.md - minimal architecture direction
  • docs/data-and-model-principles.md - model abstraction and local data principles

Contributing

Contributions should preserve the product-first, local-first, and public-safe direction of Ducelis Open.

Before contributing:

  1. read the docs in /docs
  2. avoid adding speculative enterprise or hosted features
  3. avoid hard-coding the product to a single model vendor assumption
  4. never commit real user data

License

License and third-party dependency guidance should be documented explicitly in the repository before public release.

This repository is intended to cover original Ducelis Open materials only. Third-party runtimes, models, and package dependencies remain governed by their own upstream licenses and terms.

About

Ducelis Open is the local-first AI rehearsal for high-stakes conversations

Resources

Stars

1 star

Watchers

0 watching

Forks

Contributors

Languages