Giggle
A self-hosted, customizable and ad-free Google Search experience.
What does it do?
Giggle lets you run Google searches against allow- and block-lists in an easy way. You create lists of sites to allow or block in Google's Programmable Search Engine service and Giggle will pick them up for use.
How it works
After installation, Giggle logs into Google for you to gather a list of your Custom Search Engines. When you perform a search, you select the engine to use from a dropdown and Giggle uses your Developer API key to execute a search against it. Everything is self-hosted so your credentials and data belong to you.
How it looks
Here's a glimpse of the homepage, it may look familiar 🙂
Your results contain no ads and are restricted to the sites you configure in the Programmable Search Engine console.
Getting Started
Giggle is self-hosted meaning you may host it any way you like or just run it locally. It uses NextJS so development and deployments are super easy.
What you'll need
- Some technical knowledge - I'll walk you through it in this README. If you don't already have them, you'll need NodeJS and yarn package manager.
- A Google account without MFA - You'll need to inject its username and password as environment variables. If you're curious/concerned about how they're used, check out this file. Basically, Giggle uses Puppeteer to log in to Google in order to retrieve your custom search engine IDs. If you do not have an account without MFA, just go ahead and make a new one - ⚠️ MAKE SURE NOT TO TURN ON MFA! ⚠️
- A Google Developer API key with Custom Search API privileges - Giggle uses your API key to run queries against your search engines to retrieve JSON results without ads. Head over to the Developer Console to set this up. Each engine gets 100 queries per day for free 💰
- At least one custom search engine configured in the Google account's Programmable Search Engine console - Your console should look something like this:
Setup
- Open a new Terminal window
- Clone this repository with
git clone https://github.com/dan-lovelace/giggle.git
- Change into the new directory with something like
cd giggle
- Create an empty
.env
file withtouch .env
- Open the
.env
file and paste in the following template, then update the values with your own:
GOOGLE_USERNAME=[your_email] GOOGLE_PASSWORD=[your_password] GOOGLE_API_KEY=[your_api_key]
- Run
yarn
to install dependencies - Run
yarn build
to create a new build - Run
yarn start
and wait for it to change to theready
state - Navigate to
http://localhost:3000
in a browser - If everything went right, you should see the Giggle homepage
Troubleshooting
- If you made a new Google account, it's possible the first couple login attempts by Giggle will fail due to being prompted to confirm a phone number. Try refreshing the page. You may also check the Puppeteer output in
puppeteer/screenshots
for any insight. - If you see an error saying something like "No engines found," make sure at least one search engine exists in your Programmable Search Engine console.
- Feel free to open a GitHub issue if you need additional help.
Development
Run yarn dev
to get started using the dev server.
Note: yarn dev
starts a local instance without API caching which may run up your search request limits rather quickly. There is some mock implementation in /pages/api/search.ts which will avoid hitting your engine's API during development. Uncomment the MOCK USAGE
section and comment the REAL USAGE
to use it. Mocking the search API uses the JSON located in mocks/results-recipes.json.
Testing
Giggle's tests are written using Cypress and can be run with either of these commands:
yarn cypress:open
- Opens a Cypress runner for testing individual spec files.yarn cypress:run
- Executes all tests in series.
Contribution
If you'd like to contribute, feel free to open a PR any time. Please abide by these guidelines:
- Add new tests as necessary
- Make sure the tests pass with
yarn cypress:run
- Include a description about the changes and why you think they're necessary
What's Next?
I'd like to build some sort of annotation management system to avoid having to manually create/update individual engines in the web console. It would be reduced to creating a single engine and Giggle would swap out annotation files on-the-fly. It would be much more ideal since you'd never need to visit Google's console again but it could turn out to be painfully slow. Let me know in GitHub issues if this interests you.
Also in the cards could be a fully-managed service including hosting. Users would need to hand over their own Google login credentials and API key which I can understand being hesitant about. If that's something you're interested in, please open a GitHub issue and start the conversation.
The mobile experience could be better.
from Hacker News https://ift.tt/jHmcaRt
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.