https://github.com/kern/supercharged
🏃♂️ Natural language activity tracking with GPT-3
https://github.com/kern/supercharged
activity-tracking gpt3 serverless
Last synced: about 1 month ago
JSON representation
🏃♂️ Natural language activity tracking with GPT-3
- Host: GitHub
- URL: https://github.com/kern/supercharged
- Owner: kern
- License: bsd-3-clause
- Created: 2020-08-04T02:41:41.000Z (almost 5 years ago)
- Default Branch: main
- Last Pushed: 2023-01-09T23:03:27.000Z (over 2 years ago)
- Last Synced: 2025-03-26T13:05:24.625Z (about 2 months ago)
- Topics: activity-tracking, gpt3, serverless
- Language: TypeScript
- Homepage:
- Size: 2.05 MB
- Stars: 13
- Watchers: 3
- Forks: 1
- Open Issues: 16
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# *Supercharged!*
Natural language activity tracking with GPT-3
## Deployment
1. Copy `config.example.yaml` to `config.yaml` and update it accordingly with:
- Your GPT3 API key (closed beta right now),
- Your Airtable base and table information
2. Deploy this repository as a Google Cloud function:
```shell
./scripts/deploy.sh YOUR_FUNCTION_NAME
```
3. Create a phone number in Twilio with SMS capabilities and configure its webhook URL to `POST` to the newly created Google Cloud function.
4. Text the new number with your activities in natural language and watch as your Airtable is updated automatically!
## Training
If the model makes a mistake, you can train it to learn what you mean. Update the training data in Airtable to reflect the correct labeling of each part-of-speech.
*More training resources coming soon...*
## Development
```shell
npm install
npm test
npm run build
npm run start
```
## License
BSD 3-Clause