REEN Documentation

REEN is a technology platform for scientific research and engineering development. The platform has three sections: Radar (scientific paper feed), Library (document analysis with AI-generated knowledge graphs), and Engineering (Gantt planning, AI Conference, Ex-Help). AI Conferences enable cross-AI and user multi-party conversations powered by local model subscriptions.

API Documentation

Version: 2.3.0 | Last Updated: 2026-03-05

MCP Server (AI Agents)

REEN provides a native MCP (Model Context Protocol) server that gives AI agents direct access to plans, tasks, and progress tracking.

Installation

# Clone and build
git clone https://github.com/rhoe-llc-fz/reen-mcp-server.git
cd reen-mcp-server
npm install && npm run build

Configuration (Claude Code)

Add to .mcp.json in your project root:

{
  "mcpServers": {
    "reen": {
      "command": "node",
      "args": ["/path/to/reen/mcp-server/dist/index.js"],
      "env": {
        "REEN_API_TOKEN": "reen_YOUR_TOKEN_HERE"
      }
    }
  }
}

Available Tools

Environment Variables

Example Usage

Once configured, AI agents can use REEN tools natively:

# In Claude Code conversation:
"List my plans"          → calls list_plans
"Create a new plan"      → calls create_plan
"Mark task X as done"    → calls update_task with status: "done"
"What's my progress?"    → calls get_plan_progress

Authentication

Get your API token via REEN web interface:

1. Open https://reen.tech
2. Login → User Menu → "API Tokens"
3. Create a new token and save it securely

Using the token:

curl -H "Authorization: Bearer reen_YOUR_TOKEN_HERE" \
     https://backend.reen.tech/api/gant/plans

Projects API

Project is a container for plans. Structure: {access}/{username}/{project_name}

Get Projects List

GET /api/gant/projects

Response:

{
  "projects": [
    {
      "access": "private",
      "username": "john.smith",
      "project_name": "my-project",
      "created_at": "2026-01-31T10:00:00Z"
    }
  ]
}

Create Project

POST /api/gant/projects

Request Body:

{
  "access": "private",
  "project_name": "my-new-project",
  "description": "Project description (optional)"
}

Response:

{
  "success": true,
  "project_path": "private/john.smith/my-new-project"
}

Delete Project

DELETE /api/gant/projects/{access}/{username}/{project_name}

Plans API

Plan is a structured set of tasks and subtasks following the Gant protocol.

Get Plans List

GET /api/gant/plans?project_path={project_path}

Query Parameters:

• project_path (optional) — filter by project

Create Plan

POST /api/gant/plans

Minimal Request Body:

{
  "project_path": "private/john.smith/my-project",
  "title": "Q1 Development Plan",
  "description": "Development plan for Q1 2026",
  "phases": [
    {
      "id": "phase-1",
      "title": "Phase 1: Setup",
      "description": "Infrastructure setup",
      "status": "planned",
      "subtasks": [
        {
          "title": "Setup CI/CD pipeline",
          "description": "Configure GitHub Actions",
          "status": "planned"
        }
      ]
    }
  ]
}

"briefing": "What: CI/CD setup\nHow: GitHub Actions config\nWhy: Automated deployment"

Get Plan Timeline

GET /api/gant/timeline/{plan_id}

Get Task Progress

GET /api/gant/progress/{plan_id}/{task_id}

Response:

{
  "task_id": "phase-1",
  "status": "in-progress",
  "progress": 0.5,
  "notes": [
    "[2026-01-31 10:00] phase-1: 30% — Started CI/CD setup",
    "[2026-01-31 14:00] phase-1: 50% — Docker config ready"
  ]
}

Delete Plan

DELETE /api/gant/plans/{plan_id}

Tasks API

Create Task

POST /api/gant/task

Request Body:

{
  "plan_id": "plan-2026-01-31-abc123",
  "title": "Configure Docker",
  "start_date": "2026-02-01",
  "end_date": "2026-02-15",
  "status": "planned",
  "position": 0
}

position is optional (0-based index). Omit to append at the end. If specified, existing tasks at that position and below are shifted.

Create Subtask

POST /api/gant/subtask

Request Body:

{
  "plan_id": "plan-2026-01-31-abc123",
  "task_id": "task-abc12345",
  "title": "Write Dockerfile",
  "start_date": "2026-02-01",
  "end_date": "2026-02-05",
  "status": "planned",
  "position": 0
}

Update Task

PATCH /api/gant/tasks/{task_id}

Request Body (all fields optional):

{
  "title": "New title",
  "description": "Updated description",
  "status": "in-progress",
  "progress": 0.5,
  "position": 2,
  "change_reason": "Why this change was made (recorded in audit log)",
  "change_evidence": ["ref-1", "ref-2"]
}

Valid statuses: planned, in-progress, done, blocked, cancelled. All changes are automatically recorded in the audit log via PostgreSQL triggers.

Plan Events (Audit Log)

GET /api/gant/plans/{plan_id}/events

Returns chronological audit log of all plan and task changes. Events are automatically generated by PostgreSQL AFTER UPDATE triggers — no manual logging needed.

Query Parameters:

• task_id — filter by specific task
• event_type — status_change or field_change
• field — filter by field name (title, status, description, etc.)
• actor — filter by author
• limit (default 200), offset (default 0)

Response:

{
  "events": [
    {
      "id": "evt_abc123",
      "plan_id": "argus-...",
      "task_id": "task-...",
      "entity_type": "task",
      "event_type": "status_change",
      "field": "status",
      "old_value": "planned",
      "new_value": "in-progress",
      "actor": "john.smith",
      "actor_type": "human",
      "reason": "Starting work",
      "task_title": "Task Name",
      "created_at": "2026-03-05T12:00:00Z"
    }
  ],
  "count": 42
}

Reorder Task

POST /api/gant/tasks/reorder

Move a task to a new position within its sibling group (same parent). Sibling tasks are shifted automatically.

Request Body:

{
  "task_id": "task-abc12345",
  "position": 0
}

Response:

{
  "success": true,
  "task": { "id": "task-abc12345", "position": 0, ... }
}

Gant Execution Mode Protocol

When executing a plan through an AI agent (Claude Code, GPT-5 Pro):

Notes Format

[YYYY-MM-DD HH:mm] task_id: N% — Brief description

Example:

[2026-01-31 10:00] phase-1: 30% — Dockerfile created
[2026-01-31 11:00] phase-1: 50% — Docker compose ready
[2026-01-31 12:00] phase-1: 100% — CI/CD pipeline working

Task Statuses

• planned — Scheduled
• in-progress — In progress
• done — Completed
• blocked — Blocked

Progress Values

• 0.0 — 0% (start)
• 0.3 — 30% (first milestone)
• 0.5 — 50% (half done)
• 1.0 — 100% (completed)

Usage Examples

1. Create Project and Plan

# Step 1: Create project
curl -X POST \
     -H "Authorization: Bearer reen_YOUR_TOKEN" \
     -H "Content-Type: application/json" \
     -d '{"access":"private","project_name":"q1-2026"}' \
     https://backend.reen.tech/api/gant/projects

# Step 2: Create plan
curl -X POST \
     -H "Authorization: Bearer reen_YOUR_TOKEN" \
     -H "Content-Type: application/json" \
     -d '{
       "project_path": "private/john.smith/q1-2026",
       "title": "Q1 Development",
       "phases": [{
         "id": "setup",
         "title": "Setup Infrastructure",
         "status": "planned",
         "subtasks": [{
           "title": "Configure CI/CD",
           "status": "planned"
         }]
       }]
     }' \
     https://backend.reen.tech/api/gant/plans

2. Track Progress

# Get plan timeline
curl -H "Authorization: Bearer reen_YOUR_TOKEN" \
     https://backend.reen.tech/api/gant/timeline/plan-2026-01-31-abc123

# Get specific task progress
curl -H "Authorization: Bearer reen_YOUR_TOKEN" \
     https://backend.reen.tech/api/gant/progress/plan-2026-01-31-abc123/phase-1

Narrative API

Each plan can have a narrative — a free-form text (Markdown) describing the plan's context, goals, and notes.

Get Narrative

GET /api/gant/plans/{plan_id}/narrative

Response:

{
  "narrative": "# Project Overview\n\nThis plan covers..."
}

Update Narrative

PUT /api/gant/plans/{plan_id}/narrative

{
  "narrative": "# Updated narrative\n\nNew content here..."
}

Response:

{
  "success": true,
  "narrative": "# Updated narrative\n\nNew content here..."
}

Ex-Help API

Ex-Help provides AI consultation requests linked to plans. Create a request describing a problem, generate a context pack, and paste back the AI answer.

List Ex-Help Requests

GET /api/gant/exhelp/{plan_id}

Response:

{
  "exhelp": [
    {
      "id": "exh_abc123",
      "title": "Layout issue in dashboard",
      "problem": "The sidebar overlaps...",
      "answer": "",
      "status": "draft",
      "created_at": "2026-02-15T10:00:00Z"
    }
  ]
}

Create Ex-Help Request

POST /api/gant/exhelp/{plan_id}

{
  "title": "New request",
  "problem": "Describe the problem here..."
}

Update Ex-Help Request

PATCH /api/gant/exhelp/{exhelp_id}

{
  "title": "Updated title",
  "problem": "Updated problem description",
  "answer": "AI response pasted here",
  "initial_prompt": "System prompt for external AI...",
  "status": "answered"
}

Initial Prompt

The initial_prompt field provides system-level instructions for external AI. When a share link is generated, this field becomes the primary "# Task" section in the public pack.

Statuses

• draft — Initial state, being composed
• sent — Sent for AI consultation
• answered — AI answer received

Generate Context Pack

GET /api/gant/exhelp/{exhelp_id}/pack?format=md

Generates a context pack containing the problem description, plan tasks, and relevant files. Available in md (Markdown) or json format.

Share Link (Public Access)

Generate a time-limited public URL for sharing an Ex-Help request with external AI agents (no authentication required).

POST /api/gant/exhelp/{exhelp_id}/share

Response:

{
  "share_url": "https://backend.reen.tech/api/public/exhelp/TOKEN",
  "expires_at": "2026-02-23T10:00:00Z"
}

GET /api/public/exhelp/{share_token}

Returns a Markdown document optimized for AI consumption:

1. # INSTRUCTION — directive for AI: "Execute immediately. Do NOT ask clarifying questions."
2. ## Problem — problem description
3. ## Plan — plan context with task tree
4. ## Attached Files — text files (md, json, txt, yaml, csv ≤100KB) are inlined as code blocks; binary files show metadata only
5. ## Previous Answer — prior AI answer (if any)

Artifacts API

Artifacts provides note and file storage per plan. Store research notes, implementation details, and reference files as the plan progresses.

List Artifacts

GET /api/gant/artifacts/{plan_id}

Response:

{
  "artifacts": [
    {
      "id": "art_abc123def456",
      "plan_id": "argus-20260218-...",
      "title": "Research notes",
      "content": "## Key findings\n...",
      "file_count": 2,
      "created_at": "2026-02-18T10:00:00Z",
      "updated_at": "2026-02-18T12:00:00Z"
    }
  ],
  "count": 1
}

Create Artifact

POST /api/gant/artifacts/{plan_id}

{
  "title": "New Artifact",
  "content": ""
}

Update Artifact

PATCH /api/gant/artifacts/{artifact_id}

{
  "title": "Updated title",
  "content": "## Updated content\n..."
}

Delete Artifact

DELETE /api/gant/artifacts/{artifact_id}

Soft-deletes the artifact. Associated files are preserved but hidden.

Artifact Files

Use the standard Plan Files API with context=artifact:

POST /api/gant/plans/{plan_id}/files (multipart form: file, context=artifact, artifact_id)

GET /api/gant/plans/{plan_id}/files?context=artifact&artifact_id={id}

Strategic Plans API

Strategic Plans provide a free-form canvas for organizing multiple Gantt plans visually. Each plan appears as a draggable card on an infinite React Flow canvas, with edges (arrows) connecting related plans.

List Strategic Plans

GET /api/gant/strategic-plans

Response:

{
  "strategic_plans": [
    {
      "id": "sp-a1b2c3d4e5f6g7h8",
      "user_id": "john.smith",
      "title": "Q4 2026 Strategy",
      "description": "Long-term strategic initiatives",
      "project": "main-project",
      "viewport_x": 0,
      "viewport_y": 0,
      "viewport_zoom": 1.0,
      "created_at": "2026-02-27T10:30:00Z",
      "updated_at": "2026-02-27T15:45:00Z"
    }
  ],
  "count": 1
}

Create Strategic Plan

POST /api/gant/strategic-plans

{
  "title": "Q4 2026 Strategy",
  "description": "Long-term strategic initiatives",
  "project": "main-project"
}

project is optional. SSE event strategic_plan.created is published.

Get Strategic Plan (with Cards & Edges)

GET /api/gant/strategic-plans/{sp_id}

Returns the plan metadata, all cards (with live plan title, status, progress), and edges:

{
  "strategic_plan": { "id": "sp-...", "title": "...", ... },
  "cards": [
    {
      "id": "spc-xyz123",
      "plan_id": "p-001",
      "position_x": 0, "position_y": 0,
      "width": 280, "height": 140,
      "plan_title": "Phase 1: Research",
      "plan_status": "in-progress",
      "plan_progress": 0.45
    }
  ],
  "edges": [
    {
      "id": "spe-abc789",
      "source_card_id": "spc-xyz123",
      "target_card_id": "spc-def456",
      "label": "feeds into",
      "edge_type": "default"
    }
  ]
}

Update Strategic Plan

PATCH /api/gant/strategic-plans/{sp_id}

{
  "title": "New Title",
  "description": "Updated description",
  "project": "different-project",
  "viewport_x": 100.5,
  "viewport_y": 200.3,
  "viewport_zoom": 1.5
}

All fields are optional.

Delete Strategic Plan

DELETE /api/gant/strategic-plans/{sp_id}

Soft-deletes the strategic plan. SSE event strategic_plan.deleted is published.

Canvas Cards

Cards represent Gantt plans placed on the strategic canvas. Each plan can appear once per canvas (unique constraint).

POST /api/gant/strategic-plans/{sp_id}/cards — Add a plan card

{
  "plan_id": "p-001",
  "position_x": 0,
  "position_y": 0
}

PATCH /api/gant/strategic-plans/{sp_id}/cards/{card_id} — Update position/size

{
  "position_x": 150.5,
  "position_y": -200.3,
  "width": 300,
  "height": 160
}

DELETE /api/gant/strategic-plans/{sp_id}/cards/{card_id} — Remove card (cascading: connected edges are also removed)

Canvas Edges

Edges connect cards with typed arrows. Self-loops are prevented by a database constraint.

POST /api/gant/strategic-plans/{sp_id}/edges — Add edge

{
  "source_card_id": "spc-xyz123",
  "target_card_id": "spc-def456",
  "label": "feeds into",
  "edge_type": "feeds"
}

Edge types: default, dependency, feeds, blocks. Each type renders with a distinct color.

DELETE /api/gant/strategic-plans/{sp_id}/edges/{edge_id} — Remove edge

Bulk Update

PATCH /api/gant/strategic-plans/{sp_id}/bulk-update — Update viewport and multiple card positions in a single request

{
  "viewport_x": 150.5,
  "viewport_y": -200.3,
  "viewport_zoom": 1.2,
  "cards": [
    { "id": "spc-xyz123", "position_x": 0, "position_y": 50 },
    { "id": "spc-def456", "position_x": 350, "position_y": 50 }
  ]
}

Research API (Library)

Upload documents and build interactive knowledge graphs. The pipeline extracts text, segments it into chapters, then pauses at segments_ready for your local AI to analyze via MCP.

Upload Document

POST /api/research/books/json

{
  "title": "Introduction to TRIZ",
  "text": "Full text content of the document...",
  "author": "G. Altshuller",
  "domain": "engineering",
  "language": "ru"
}

Response:

{
  "id": "rb-abc123",
  "status": "pending",
  "message": "Book uploaded, processing started"
}

Also supports multipart form upload via POST /api/research/books (accepts PDF files).

List Documents

GET /api/research/books

Query params: ?domain=, ?status=

Get Document with Knowledge Graph

GET /api/research/books/{book_id}

Returns the document metadata, all cards (one per chapter), and edges (relationships between chapters). Only meaningful when processing_status = "completed".

Check Processing Status

GET /api/research/books/{book_id}/status

Response:

{
  "processing_status": "segments_ready",
  "processing_stage": "segmenting",
  "processing_progress": 1.0,
  "total_chapters": 12
}

Processing Statuses

Get Segments (for Local AI)

GET /api/research/books/{book_id}/segments

Only works when status is segments_ready. Returns raw text segments and the expected JSON schema for analysis output.

Response:

{
  "book_id": "rb-abc123",
  "title": "Introduction to TRIZ",
  "language": "ru",
  "total_segments": 12,
  "segments": [
    { "heading": "Chapter 1: Inventive Problems", "text": "...", "start_line": 1 },
    { "heading": "Chapter 2: Contradictions", "text": "...", "start_line": 45 }
  ],
  "expected_output_schema": { ... }
}

Submit Analysis

POST /api/research/books/{book_id}/analyze

{
  "book_summary": "A foundational text on systematic invention methodology...",
  "cards": [
    {
      "chapter_number": 1,
      "title": "Inventive Problems",
      "essence": "Inventive problems contain contradictions that cannot be resolved by compromise",
      "summary_simple": "Introduces the concept of inventive problems...",
      "summary_technical": "Defines inventive problems as those containing technical or physical contradictions...",
      "importance": 5,
      "key_terms": ["inventive problem", "contradiction", "compromise"],
      "evidence_quotes": [{ "text": "An inventive problem arises when...", "location": "p.12" }]
    }
  ],
  "edges": [
    {
      "from_chapter": 1,
      "to_chapter": 2,
      "type": "depends_on",
      "confidence": 0.95,
      "why": "Chapter 2 builds on the contradiction concept from Chapter 1"
    }
  ]
}

Edge types: depends_on, extends, illustrates

Card Position (Drag & Save)

PATCH /api/research/cards/{card_id}/position

{ "x": 250.5, "y": 100.0 }

Edge Management

POST /api/research/edges — Create a manual edge

PATCH /api/research/edges/{edge_id}/verify — Verify or reject an edge

DELETE /api/research/edges/{edge_id} — Delete an edge

Delete Document

DELETE /api/research/books/{book_id}

Conferences API

AI Conference provides a messenger-style chat for multi-party conversations between users and AI models (Claude, GPT, Gemini, Grok).

List Conferences

GET /api/conferences

Response:

{
  "conferences": [
    {
      "id": "conf_b1e548af2763",
      "title": "Architecture Review",
      "owner": "john.smith",
      "created_at": "2026-02-08T10:00:00Z",
      "message_count": 15,
      "agents": ["reen-cli-user"]
    }
  ]
}

Create Conference

POST /api/conferences

{
  "title": "My Conference",
  "description": "Optional description"
}

Get Conference with Messages

GET /api/conferences/{id}?limit=100

Returns conference metadata and the last N messages.

Delete Conference

DELETE /api/conferences/{id}

Initial Prompt

Each conference can have a system prompt that sets the context for all AI models.

GET /api/conferences/{id}/initial-prompt

Response:

{
  "initial_prompt": "You are an expert architect..."
}

PUT /api/conferences/{id}/initial-prompt

{
  "initial_prompt": "You are an expert architect reviewing our system design..."
}

WebSocket

Connect to a conference in real-time:

wss://backend.reen.tech/ws/conference/{id}?token=reen_YOUR_TOKEN

Send a message:

{"type": "message", "content": "@claude explain this code", "mentions": ["claude"]}

Receive messages:

{"type": "message", "id": "msg_abc", "author": "claude", "role": "assistant", "content": "...", "ts": "..."}

@mention Routing

Conference Control (Go / Stop)

The UI provides a Go/Stop toggle to control AI conversation flow:

WebSocket control messages:

{"type": "control", "action": "playing"}   // Resume (Go)
{"type": "control", "action": "stopped"}   // Stop + cancel all generations

Model-to-Model Conversations

Models respond sequentially within each round: Claude → GPT → Gemini → Grok. Each model sees the responses from previous models in the same round.

Round Chaining (v1.5)

When a model uses @mention in its response, it triggers a follow-up round automatically:

1. Round 1: User sends @all discuss X → Claude responds → GPT sees Claude's answer → Gemini sees both → Grok sees all
2. Round 2: If responses contain @mentions, mentioned models respond (with full context)
3. Rounds 3–5: Chain continues while @mentions exist
4. Auto-pause: When no more @mentions or max 5 rounds reached

@mention Rules

Models are instructed to end each response with a direct question using @ to keep the conversation going.

Example flow:

User: "@all discuss the future of AI"

Round 1 (automatic):
  → Claude responds (sees user message)
  → GPT responds (sees user + Claude)
  → Gemini responds (sees user + Claude + GPT)
  → Grok responds (sees all above)

Round 2 (triggered by @mentions in Round 1):
  → @mentioned models respond with full context

... up to 5 rounds, then auto-pause

File Content in Prompts

When users upload text files (.md, .txt, .json, .py, .js, etc.) to a conference, the file content is included in the model's prompt context (up to 16KB per file). Models can analyze and discuss uploaded files.

Send Message via REST

POST /api/conferences/{id}/messages

{
  "content": "Hello @claude, analyze this",
  "author": "claude-code",
  "mentions": ["claude"]
}

Allows external agents (Claude Code, scripts) to post messages into conferences.

reen-cli

Connect local AI models to conferences:

npm install -g reen-cli

# Connect to all conferences (recommended)
reen-cli daemon --token reen_YOUR_TOKEN

# Connect to a single conference
reen-cli connect conf_XXXX --token reen_YOUR_TOKEN

# List conferences
reen-cli list --token reen_YOUR_TOKEN

Source: github.com/rhoe-llc-fz/reen-cli

Error Codes

Rate Limits

Default for API token: 60 requests/minute

When exceeded:

{
  "error": "Rate limit exceeded",
  "retry_after": 30
}

Support

• API Documentation: https://reen.tech/documentation
• X (Twitter): @reentechnology
• Email: [email protected]

REEN Platform API v2.3.0 | © 2026 RHOE LLC FZ