CLEAgora

Community-based qualitative research platform built on Discourse.

CLEAgora adapts the CLEA methodology (Consumer-Level Elicitation Architecture) — originally designed for 1:1 AI-guided conversations — into a collective, forum-based format where recruited respondents interact with each other in closed communities.

Part of the DECODia ecosystem by [h2](https://h2.fr).

Concept

• Closed communities: respondents are invited by email into a private Discourse instance configured for a specific study.
• Collective-first: the value comes from debate, co-construction, and organic interaction between participants.
• Flash deployment: brief → auto-config → live community → restructured output in 72h.
• Structured output: an extraction + enrichment pipeline transforms forum content into analysis-ready material (segmented verbatims, thematic tags, interaction graphs, Markdown write-ups).

Status

Exploration stage — validating the concept. No application code yet. Design documents and reference material are in context/.

Project structure

CLAUDE.md              # AI assistant instructions
DEVS.md                # Developer workflow guide
context/
  vision.md            # Purpose, users, constraints, stage
  journal.md           # Session log (reverse chronological)
  decisions.md         # Architectural/product decisions
  backlog.md           # Now / Next / Later tasks
  tech-stack.md        # Technologies and conventions
  references/          # Design docs, research, meeting notes

Getting started

Setup instructions will be added once the Discourse instance and Python pipeline are in place.

See DEVS.md for the developer workflow and context/backlog.md for current priorities.

License

Proprietary — h2\ / Ubyx.