Build a Team Knowledge Base

The scenario

A team of three devs working on a payments service:

Alex just watched a great Stripe webhooks tutorial
Jordan wrote architecture notes about the retry system
Sam is new and needs to get up to speed

They share one ChatGrid board. Here’s how it works.

Alex adds a tutorial

Alex found a YouTube video explaining Stripe webhook best practices.

curl -X POST https://api.chatgrid.ai/v1/boards/{boardId}/nodes \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "youtube",
    "position": {"x": 0, "y": 0},
    "data": {"url": "https://youtube.com/watch?v=stripe-webhooks-101", "label": "Stripe webhooks tutorial"}
  }'

curl -X POST https://api.chatgrid.ai/v1/boards/{boardId}/documents/vectorize \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "content": "Stripe webhook tutorial notes: verify signatures, make handlers idempotent, retry transient failures, and move long-running work to a background queue.",
    "node_id": "{tutorialNodeId}",
    "metadata": {"source": "stripe-webhooks-101", "type": "video-notes"}
  }'

> Add this video to the payments board: https://youtube.com/watch?v=stripe-webhooks-101

The node keeps the tutorial visible on the board. The vectorized notes make it searchable for every team member and their AI tools.

Jordan adds architecture notes

Jordan documents how the retry system works.

curl -X POST https://api.chatgrid.ai/v1/boards/{boardId}/documents/vectorize \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "content": "## Retry System\n\nFailed webhooks go through Inngest with exponential backoff.\nMax 5 retries over 24 hours.\nAfter exhaustion, events land in the dead letter queue (Supabase table: failed_webhooks).\nAlert fires in Sentry if DLQ depth > 10.\n\nConfig: src/server/inngest/functions/stripeWebhook.ts",
    "metadata": {"source": "architecture-notes", "author": "jordan"}
  }'

> Add these notes to the payments board:
  "## Retry System
  Failed webhooks go through Inngest with exponential backoff.
  Max 5 retries over 24 hours. After exhaustion, events land in
  the dead letter queue (Supabase table: failed_webhooks).
  Alert fires in Sentry if DLQ depth > 10.
  Config: src/server/inngest/functions/stripeWebhook.ts"

Sam searches the board

Sam just joined. They need to understand how webhook retries work.

curl -X POST https://api.chatgrid.ai/v1/boards/{boardId}/documents/search \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "how do webhook retries work?", "limit": 5}'

> Search the payments board for webhook retry logic

The search returns Jordan’s architecture notes AND relevant parts of Alex’s YouTube video. Sam gets the full picture without bothering anyone.

Sam asks a follow-up question

Claude Code

> Based on our payments board, what happens when a webhook fails
  after all retries are exhausted?

AI responds using the board’s actual content:

Based on Jordan's architecture notes:
Failed webhooks go to the dead letter queue (Supabase table: failed_webhooks).
A Sentry alert fires if the DLQ depth exceeds 10.
The config is in src/server/inngest/functions/stripeWebhook.ts.

No guessing. No hallucination. Grounded in what the team actually documented.

Why this matters

If Alex had shared that tutorial in Slack, it would be gone in 2 days. If Jordan had put those notes in a Google Doc, Sam would never find it.On the board, both are searchable forever. By humans and AI.

Without ChatGrid	With ChatGrid
Knowledge dies in Slack	Knowledge lives on the board
New devs bother senior devs	New devs ask AI, get sourced answers
Tribal knowledge leaves with people	Knowledge stays with the project
AI gives generic answers	AI gives answers from your actual docs

Cookbooks

Build a Team Knowledge Base

The scenario

Alex adds a tutorial

Jordan adds architecture notes

Sam searches the board

Sam asks a follow-up question

Why this matters

Next steps

Add Videos & PDFs

MCP Workflow

​The scenario

​Alex adds a tutorial

​Jordan adds architecture notes

​Sam searches the board

​Sam asks a follow-up question

​Why this matters

​Next steps

Add Videos & PDFs

MCP Workflow

The scenario

Alex adds a tutorial

Jordan adds architecture notes

Sam searches the board

Sam asks a follow-up question

Why this matters

Next steps