MVP Scope — What Was Delivered

Overview

The Yappa Knowledge Hub MVP is a complete, working system that enables teams to capture, organize, and digest knowledge through Slack with AI-generated summaries and Notion integration.

Core Features

1. Knowledge Capture via Slack

Users capture knowledge without leaving Slack:

Message Shortcut — Save any Slack message as knowledge via the three-dot menu
Global Shortcut — Quick add from anywhere in Slack
App Home Button — Add knowledge from the dashboard
URL Content Extraction — Automatic page content extraction when a URL is included
Auto-sync to Notion — Every knowledge item creates a Notion page

2. Category Management

Organise knowledge into thematic lists:

Create categories with name, description, and emoji icon
Categories synced to Notion bidirectionally
Target groups (audience labels) synced from Notion multi-select
Digest configuration per category (frequency, day, time)

3. AI-Powered Summaries

Automatic Dutch-language summaries for knowledge items:

2–3 sentence summaries generated by OpenAI or OpenRouter
Lazy generation — summaries created when needed for digests
Stored in Knowledge.highlight field
Graceful fallback to content truncation on AI failure
Single hardcoded prompt optimised for Dutch summaries

4. Digest Generation

AI-enhanced digest reports per category:

7-step pipeline: category → date range → query → AI highlights → format → persist → Notion sync
Supports multiple date modes: last N days, date range, all time
Dutch-language Markdown formatting with per-item summaries, tags, and read time
Immutable snapshots — digests preserve the AI output at generation time
Synced to Notion with full content embedding

5. Digest Distribution

Deliver digests to the right people:

Slack DM delivery to category subscribers
Per-user delivery tracking with retry support (max 3 attempts)
Automatic subscriber resolution + initiator inclusion
Markdown → Slack mrkdwn conversion
Distribution results synced to Notion

6. Subscriptions

User-controlled notification preferences:

Subscribe/unsubscribe to categories via App Home overflow menu
Soft-delete with history preservation
Unique constraint: one subscription per user per category

7. Notion Integration

Bidirectional sync for all data:

4 Notion databases: Knowledge, Categories, Digests, Subscriptions
Real-time outbound sync on creation/update
On-demand inbound sync from Notion
Full Markdown content embedding in Notion digest pages
Dual-ID resolution (local integer + Notion UUID)

8. Digest Scheduling

Automated digest generation:

Per-category scheduling: frequency (weekly/biweekly/monthly/quarterly)
Day of week and time configuration
Scheduler service checks due categories and triggers generation

Technical Architecture

Database

Aspect	Implementation
Engine	SQLite
ORM	Doctrine ORM with attribute mapping
Entities	5: Category, Knowledge, Digest, DigestDelivery, Subscription
Migrations	Schema created via `doctrine:schema:create`
Timestamps	Gedmo Timestampable (auto-managed createdAt/updatedAt)

Service Architecture

Controller → Orchestrator → Domain Services → Persistence → Repository
                                    ↕
                              AI Provider
                                    ↕
                            Notion Sync Service

38 service files organized by domain:

Orchestrator layer (2 files) — coordinates multi-step workflows
Digest domain (10 files) — generation, formatting, distribution, scheduling, delivery tracking
Knowledge domain (3 files) — creation, persistence, query
Subscription domain (3 files) — facade + CQRS split
Notion domain (11 files) — sync, repositories, page builder, mappers
AI domain (2 files) — provider interface + OpenAI/OpenRouter implementation
Content domain (2 files) — extraction + HTML parser
Transport (2 files) — message transport interface + Slack implementation
Root (2 files) — CategoryService + EmojiConverter

API Surface

18 authenticated endpoints across 6 controllers:

Controller	Endpoints
CategoryController	POST/GET /api/categories
CategorySettingsController	GET/PUT /api/categories/{id}/settings/digest
DigestController	POST generate, POST distribute, GET delivery-status, POST retry-failed, GET by-category
KnowledgeController	POST/GET /api/knowledge
NotionSyncController	POST from-notion, POST categories-from-notion, POST to-notion
SubscriptionController	POST subscribe, POST unsubscribe, GET user/{userId}, GET category/{categoryId}/subscribers

Authentication

All API endpoints require ROLE_ADMIN via bearer token authentication. The AccessTokenHandler validates tokens configured via environment variables.

AI Provider

Configuration via environment variables: OPENAI_MODEL, OPENAI_API_KEY, OPENAI_API_URL.

Entities

Knowledge

Field	Type	Description
id	int (auto)	Primary key
title	string (255)	Title (3-255 chars, required)
content	text	Main content + extracted URL content
tags	JSON	User-defined tags
targetGroups	JSON	Audience labels
userId	string	Slack user ID of creator
status	string	pending/Not started/In progress/Done/approved/rejected/archived/draft/published
url	string (500)	Source URL
urlMetadata	JSON	Extraction details
sourceMessage	JSON	Original Slack message data
highlight	text	AI-generated 2-3 sentence Dutch summary
summaries	JSON	AI summary data from Notion sync
notionId / notionUrl	string	Notion references
createdAt/updatedAt	datetime	Gedmo timestamps

Digest

Field	Type	Description
id	int (auto)	Primary key
title	string (255)	Auto-generated title
content	text	Full Markdown digest
period	string	Date range label
status	string	Done/sent
statistics	JSON	highlightCount, days, tokenCount, cost, durationMs
knowledgeHighlights	JSON	Immutable snapshot of highlights at generation time
generatedBy	string	Slack user who triggered generation
notionId / notionUrl	string	Notion references
createdAt/updatedAt	datetime	Gedmo timestamps
category	ManyToOne	Required category reference
knowledges	ManyToMany	Linked knowledge items

DigestDelivery

Field	Type	Description
id	int (auto)	Primary key
userId	string	Recipient Slack user ID
deliveryChannel	string	slack/notion/email
status	string	pending/sent/failed
retryCount	int	Retry attempts (max 3)
errorMessage	text	Failure details
sentAt	datetime	Delivery timestamp
digest	ManyToOne	Required digest reference (CASCADE)

Subscription

Field	Type	Description
id	int (auto)	Primary key
userId	string	Subscriber Slack user ID
slackUserName	string	Display name
isActive	bool	Default true
subscribedAt	datetime	Subscription timestamp
unsubscribedAt	datetime	Unsubscription timestamp (null if active)
notionId	string	Notion page reference
category	ManyToOne	Required category reference (CASCADE)

What's Not in the MVP

Feature	Status	Notes
Per-role AI summaries	Not built	Single summary per item
Prompt template management	Not built	Hardcoded prompt constant
Knowledge edit/delete UI	Not built	Only create + list
Advanced search	Not built	Basic keyword search only
PDF/RSS/YouTube processing	Not built	HTML only
Redis caching	Not built	Not needed for current scale
Symfony Messenger	Not built	Synchronous processing
User authentication system	Not built	Simple bearer token
Web dashboard	Not built	Slack-only interface

MVP Scope — What Was Delivered ​

Overview ​

Core Features ​

1. Knowledge Capture via Slack ​

2. Category Management ​

3. AI-Powered Summaries ​

4. Digest Generation ​

5. Digest Distribution ​

6. Subscriptions ​

7. Notion Integration ​

8. Digest Scheduling ​

Technical Architecture ​

Database ​

Service Architecture ​

API Surface ​

Authentication ​

AI Provider ​

Entities ​

Category ​

Knowledge ​

Digest ​

DigestDelivery ​

Subscription ​

What's Not in the MVP ​