Skip to content
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 19 additions & 4 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -19,8 +19,8 @@ mid-scene. That's the whole idea. The script is the published documentation at
*"wait, how do I append an event again?"* — Prompter whispers the answer, **with citations**, or admits
honestly when the docs don't cover it.

No hallucinated APIs. No confidently-wrong snippets. Just the docs, retrieved and read back to you where you
already are.
It is built to stick to the script: answers come from the retrieved docs, with citations — and when the docs
don't cover something, it says so instead of improvising.

## 🎭 Why "Prompter"?

Expand Down Expand Up @@ -75,7 +75,7 @@ you at a human. That honesty is the feature.
requests, ideas, docs gaps — anything worth tracking.

> Prompter never barges into normal conversation — it only speaks when spoken to, and it rate-limits each
> person to a handful of questions per window so no one can spam it..
> person to a handful of questions per window so no one can spam it.

## 🧠 How it works

Expand All @@ -95,7 +95,7 @@ flowchart LR
- **Ingest** — walk cratis.io's sitemap, fetch each page's markdown mirror, strip the MDX noise, and split
into heading-aware chunks. Only changed chunks are re-embedded, so re-indexing is cheap.
- **Retrieve** — one SQL query fuses lexical (BM25 via `tsvector`) and semantic (cosine over pgvector) hits
with Reciprocal Rank Fusion. Anthropic's own benchmarks show hybrid roughly halves retrieval misses.
with Reciprocal Rank Fusion.
- **Answer** — Claude gets the top passages and a system prompt that demands grounding and citations, and
refuses when the score says the docs don't have it.

Expand Down Expand Up @@ -180,6 +180,21 @@ And one more gate that's unusual for a bot: **answer quality is measured, not vi
harness (milestone M4) scores groundedness, citation accuracy, and refusal behavior — and gates prompt and
retrieval changes the same way specs gate code.

## The Cratis ecosystem

This project is part of [Cratis](https://www.cratis.io) — free, MIT-licensed tools for building event-sourced and CQRS applications.

- **[Chronicle](https://github.com/Cratis/Chronicle)** — event-sourcing database and runtime. Orleans-based kernel, pluggable storage (MongoDB default; PostgreSQL, SQL Server, SQLite, in-memory), language-agnostic gRPC contracts. [Docs](https://www.cratis.io/chronicle/)
- **Chronicle clients** — first-class [.NET SDK](https://github.com/Cratis/Chronicle), plus [TypeScript](https://github.com/Cratis/Chronicle.TypeScript), [Kotlin/Java](https://github.com/Cratis/Chronicle.Kotlin), and [Elixir](https://github.com/Cratis/Chronicle.Elixir); [Python](https://github.com/Cratis/Chronicle.Python) coming soon (pre-alpha). AI agents connect through the [Chronicle MCP server](https://github.com/Cratis/Chronicle.Mcp).
- **[Arc](https://github.com/Cratis/Arc)** — opinionated CQRS framework for ASP.NET Core with commands, queries, validation, authorization, and TypeScript proxy generation. Works without event sourcing. [Docs](https://www.cratis.io/arc/)
- **[Components](https://github.com/Cratis/Components)** — React components aligned with Arc patterns. [Docs](https://www.cratis.io/components/)
- **[CLI](https://github.com/Cratis/cli) + Workbench** — inspect and diagnose Chronicle from the terminal or the browser. [Docs](https://www.cratis.io/cli/)
- **Model-first layer (experimental)** — [Studio](https://github.com/Cratis/Studio), [Screenplay](https://github.com/Cratis/Screenplay), [Stage](https://github.com/Cratis/Stage), [Scene](https://github.com/Cratis/Scene), [Prologue](https://github.com/Cratis/Prologue)
- **Supporting** — [Fundamentals](https://github.com/Cratis/Fundamentals), [Specifications](https://github.com/Cratis/Specifications), [Synopsis](https://github.com/Cratis/Synopsis), [Lens](https://github.com/Cratis/Lens), [Narrator](https://github.com/Cratis/Narrator), and free [AI tooling](https://github.com/Cratis/AI) (preview); [Ensemble](https://github.com/Cratis/Ensemble) coming soon (pre-release)
- **[Samples](https://github.com/Cratis/Samples)** — runnable event sourcing and CQRS samples for the whole stack

Everything Cratis publishes today is MIT licensed and free to use.

---

<div align="center">
Expand Down
Loading