Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/khoj-ai/flint
Journal with Khoj on WhatsApp. Supports multi-lingual audio messages, references previous conversations for context aware responses
https://github.com/khoj-ai/flint
ai bot personal-assistant python whatsapp
Last synced: about 4 hours ago
JSON representation
Journal with Khoj on WhatsApp. Supports multi-lingual audio messages, references previous conversations for context aware responses
- Host: GitHub
- URL: https://github.com/khoj-ai/flint
- Owner: khoj-ai
- License: gpl-3.0
- Created: 2023-08-03T03:26:40.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-11-09T17:21:16.000Z (3 months ago)
- Last Synced: 2024-11-09T18:25:33.004Z (3 months ago)
- Topics: ai, bot, personal-assistant, python, whatsapp
- Language: Python
- Homepage:
- Size: 162 KB
- Stars: 49
- Watchers: 3
- Forks: 7
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
An open, personal AI that grows with you.
[📜 Read Docs](https://docs.khoj.dev)
•
[🌍 Visit Website](https://khoj.dev)
•
[💬 Get Involved](https://discord.gg/BDgyabRM6e)
***
Flint makes Khoj, your personal AI, accessible over Whatsapp.
So you can converse with it easily, on your phone or computer.
***
## ⭐️ Chat with Khoj on WhatsApp
If you want to start chatting with our hosted Khoj right away, you have two options:
1. 🪧 Use the QR code
- Scan the QR code below using your phone's camera.
[](https://wa.me/18488004242)
2. 📞 Directly use the phone number
- Add the number [+1 (848) 800-4242](https://wa.me/18488004242) to your contacts and send a message to it on WhatsApp.
## Run it yourself
You can get setup with your own instance of Khoj via WhatsApp in a few simple steps.
### Start the service
#### Docker setup
1. Fill in the relevant environment variables in the `docker-compose.yml` file under the `app` service.
2. Start the service.
```bash
$ docker-compose up
```
##### Run migrations
In order to setup the database, you need to run the migrations. This needs to be done before you can start using the service, and anytime a new migration is added.
1. SSH into the docker container. You can get the name of the container by running `docker container ls`.
```bash
$ docker exec -it khoj_app_1 bash
```
2. Run the migrations.
```bash
$ python3 src/flint/manage.py migrate
```
#### Test that it's working
1. Go to `localhost:8488/docs` in your browser. You should see the Swagger UI.
2. Click on the `/dev/chat` endpoint.
3. Click on the "Try it out" button.
4. Enter any prompt in the `Body` field and click on the "Execute" button. You should get a response from the flint service.
If this didn't work, you might need to debug what's wrong.
### Setup Ngrok
1. Install [ngrok](https://ngrok.com/download).
2. Run ngrok on port 8488.
```bash
$ ngrok http 8488
```
3. This will output a url, like `https://abcd-ef-ghi-123-456.ngrok-free.app`. Copy this url and use it in the next step.
### Setup Twilio
1. Create a Twilio account.
2. In the [Twilio console](https://console.twilio.com/us1/develop/sms/try-it-out/whatsapp-learn), setup a Sanbox for Whatsapp.
3. Send a message to the Twilio Whatsapp number with the secret code. You should get a response first from Twilio
3. In the Sandbox Setings tab, set the webhook to the ngrok url at the `/api/chat` endpoint. From the previous example, that would be `https://abcd-ef-ghi-123-456.ngrok-free.app/api/chat` under the "WHEN A MESSAGE COMES IN" section.
4. Send a message to the Twilio Whatsapp number. You should get a response from the flint service.
### Completion
And you're done! You can now chat with your bot over Whatsapp. When you're done finalizing your bot, make sure to flip the `DEBUG` flag in the `docker-compose.yml` file to `False` or remove it altogether.
## Usage
Khoj can handle multiturn conversations and can continue the conversation from where you left off. It works well as a companion for though and reasoning, and can be used to keep track of your thoughts and ideas. You can think of it as a journal that you can talk to.
## Details
This service works by provisioning a Postgres database on your machine that stores your conversation history. Every additional number that chats with your bot will have a separate conversation history. This is done by using the Twilio `From` number as the key in the database.