Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Intelligent error monitoring agent that uses Airweave to find context from your code, tickets, and Slack. Clusters similar errors, enriches with relevant context, and creates actionable alerts.
| Date | Stars |
|---|---|
| 2026-07-24 | 367 |
| 2026-07-25 | 367 |
| 2026-07-28 | 367 |
| 2026-07-30 | 367 |
| 2026-07-31 | 368 |
| 2026-08-06 | 368 |
Today
— stars today
This week
+1 stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.27%/day
# Intelligent Error Monitoring Agent
Build production-grade error monitoring with **semantic understanding**. This project shows how [Airweave](https://airweave.ai) transforms raw errors into actionable insights by connecting your codebase, tickets, and Slack discussions.
> **Background:** This project is based on our internal agent "Donke" that handles ~40,000 Airweave queries/month. Read more about [how we use Airweave to improve Airweave](https://airweave.ai/blog/how-we-use-airweave-to-improve-airweave).
<video width="100%" src="https://github.com/user-attachments/assets/4756ef10-8a8a-440f-a397-609e8afb3e44" controls></video>
## Why This Exists
Error monitoring tools give you alerts. What you actually need is **context**:
- What code is involved?
- Has someone already worked on this?
- Is this a new issue or a regression?
This agent uses Airweave to automatically find that context and make intelligent decisions about severity, deduplication, and alerting.
## What It Does
```
Raw Errors → [Clustering] → [Context Search] → [Analysis] → Intelligent Alerts
│ │ │
└── Groups └── Airweave └── Severity + status
similar finds code, determination with
errors tickets, Slack suppression logic
```
*Example: 20 raw errors → 4 actionable clusters → 4 alerts (instead of 20)*
---
# Part 1: Interactive Demo
The frontend provides a visual demonstration of how the error monitoring pipeline works. It's designed to showcase the tool's capabilities with sample data.
## Quick Start (2 minutes)
```bash
# Clone and setup
git clone <repo-url>
cd support-agent
cp .env.example .env
# Optional: Add LLM key for smarter clustering (works without it too)
# OPENAI_API_KEY=your_key (or ANTHROPIC_API_KEY)
# Start backend
cd backend
python -m venv venv && source venv/bin/activate
pip install -r requirements.txt
uvicorn main:app --reload --port 8000
# Start frontend (new terminal)
cd frontend
npm install && npm run dev
```
Open [http://localhost:3000](http://localhost:3000) and click **Run Demo**.
## What the Demo Shows
The interactive visualization walks you through:
1. **Raw Errors** - 20 sample production errors from a data sync platform
2. **Semantic Clustering** - Errors grouped by root cause (rate limiting, DB issues, auth failures, etc.)
3. **Context Search** - Related code, Linear tickets, and Slack threads found via Airweave
4. **Severity Analysis** - S1-S4 classification with reasoning
5. **Alert Preview** - What Slack messages and Linear issues would be created
The demo uses:
- **Sample data** by default (realistic errors from a SaaS platform)
- **Mock search results** showing what Airweave would return
- **Preview mode** for Linear/Slack (no actual tickets/messages created)
---
# Part 2: Production Setup
To use this as an actual error monitoring agent for your application:
## Step 1: Connect Your Error Source
Replace sample data with real errors from your monitoring stack.
### Option A: Sentry
```bash
DATA_SOURCE=sentry
SENTRY_AUTH_TOKEN=your_token
SENTRY_ORG_SLUG=your_org
SENTRY_PROJECT_SLUG=your_project # Optional: filter to specific project
```
### Option B: Azure Log Analytics
```bash
DATA_SOURCE=azure
AZURE_TENANT_ID=your_tenant
AZURE_CLIENT_ID=your_client
AZURE_CLIENT_SECRET=your_secret
AZURE_LOG_ANALYTICS_WORKSPACE_ID=your_workspace_id
```
### Option C: Custom Source
Create `backend/sources/your_source.py`:
```python
from sources.base import DataSource
class YourSource(DataSource):
@property
def name(self) -> str:
return "Your Source"
async def fetch_errors(self, window_minutes=30, limit=100):
# Fetch from your API and normalize to RawError format
return [RawError(...) for error in your_api.get_errors()]
```
## Step 2: Connect Airweave for Context Search
This is where the magic happens - Airweave searches your GitHub code, LineaExcerpt of 10,390 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:ef4419db5c65d53a, topic:observability, topic:monitoring
matched fp:ef4419db5c65d53a, topic:semantic-search
matched fp:ef4419db5c65d53a, topic:llm