https://github.com/sette7blo/feedme
Self-hosted AI recipe manager with meal planning and pantry
https://github.com/sette7blo/feedme
docker flask meal-planning recipe-manager self-hosted sqlite
Last synced: about 2 months ago
JSON representation
Self-hosted AI recipe manager with meal planning and pantry
- Host: GitHub
- URL: https://github.com/sette7blo/feedme
- Owner: sette7blo
- License: mit
- Created: 2026-04-12T20:18:05.000Z (2 months ago)
- Default Branch: main
- Last Pushed: 2026-04-12T20:40:08.000Z (2 months ago)
- Last Synced: 2026-04-12T22:24:42.316Z (2 months ago)
- Topics: docker, flask, meal-planning, recipe-manager, self-hosted, sqlite
- Language: HTML
- Size: 91.8 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- License: LICENSE
Awesome Lists containing this project
README
# Feedme — Recipe Intelligence
A self-hosted recipe management platform. Import recipes from AI generation, RSS feeds, URLs, or photos. Plan meals, track your pantry, and generate grocery lists. Runs entirely in Docker.
---
## Features
- **AI generation** — describe a dish and get a full recipe with photo
- **RSS import** — subscribe to recipe sites; new recipes land automatically
- **URL import** — paste any recipe page URL to extract and save it
- **Image import** — photograph a cookbook page or recipe card; AI extracts it
- **Staging workflow** — all imports require your approval before going active
- **Meal planner** — assign recipes to days across a weekly calendar
- **Pantry tracking** — log what you have with quantities; scan barcodes to look up products instantly via Open Food Facts
- **Cook tonight** — filter your recipe library by what you already have in the pantry, sorted by ingredient coverage
- **Grocery list** — automatically calculates what to buy based on your meal plan and pantry
- No account, no cloud, no tracking — your data stays on your server
---
## Quick Start
No build required — pull straight from Docker Hub.
**1. Create a `docker-compose.yml`:**
```yaml
services:
feedme:
image: dockersette/feedme:latest
ports:
- "5000:5000"
volumes:
- ./recipes:/app/recipes
- ./images:/app/images
- ./data:/app/data
- ./.env:/app/.env
restart: unless-stopped
```
**2. Create a `.env` file:**
```env
PPQ_API_KEY=your-key-here
PPQ_BASE_URL=https://api.ppq.ai/v1
PPQ_MODEL=claude-haiku-4-5
PPQ_IMAGE_MODEL=gpt-image-2
PPQ_VISION_MODEL=claude-haiku-4-5
FLASK_SECRET=change-me-to-something-random
```
**3. Start it:**
```bash
docker compose up -d
```
**4. Open it:**
```
http://YOUR_SERVER_IP:5000
```
---
## Updating
```bash
docker compose pull && docker compose up -d
```
---
## AI Provider
Feedme uses any OpenAI-compatible endpoint. The recommended provider is [PPQ.ai](https://ppq.ai), which gives access to Claude and OpenAI models via a single API key and endpoint.
You can configure the key and models directly in the Settings tab after first launch. PPQ.ai offers a wide range of text, vision, and image generation models — pick what suits your budget and quality preferences.
---
## Port
Default is **5000**. Change the left side of the ports mapping to use a different host port:
```yaml
ports:
- "8080:5000" # serve on port 8080 instead
```
---
## Security
Feedme has no authentication. It is designed for **personal / home server use only**, behind a firewall or VPN. Do not expose port 5000 to the public internet without adding an auth layer (e.g. HTTP Basic Auth via an nginx reverse proxy).
---
## License
MIT