node package manager
We need your input. Help make JavaScript better: Take the 2017 JavaScript Ecosystem survey »


Gitter Archive

Gitter Chat room message downloader

npm npm


Install yarn package manager.

Install gitter-archive-cli as global.

$ yarn global gitter-archive-cli
# OR 
$ npm install -g gitter-archive-cli

Add a .gitterarchiverc[.json|.js|.yml|.yaml] or gitterarchive.config.js file in the directory where you execute the CLI in the following format.

The configuration is loaded using cosmiconfig module.
The archive and no archive room lists are matched using extglob module.

NOTE: if you're copy pasting the following example, remove the comments.

  "rooms": {
    "noArchiveList": [
      // rooms not to archive
      // e.g.
    "archiveList": [
      // list of rooms to archive
      // e.g.
  "groups": {
    "enabled": [{
      "uri": "FreeCodeCamp",
      "id": "57542cf4c43b8c6019778297"
    "disabled": [
      // groups not to be scanned

Copy the example.env as .env and fill in the details.
Start the archive process.

$ gitter-archive

Use node v7 to run.

# run build 
$ yarn run build
# start gitter archive 
$ yarn run start


$ yarn global nodemon
$ nodemon -q -x yarn run start:dev

OR for build watch

$ yarn global nodemon
$ yarn run build:watch

Then link it as global

# for globally making the cli and the module available 
$ npm link



CLI options

  • --boost - boost archiving using skip parameter
  • -b - beforeId
  • --gid - group id (required)
  • --guri - group uri (optional)
  • --uri - room uri (optional)
  • --id - room id (required)
  • --ignore, -i - ignore room uri pattern (multiple)


With a limit of 100 messages/request and 100 requests/minute,
total downloadable messages/minute = 100 x 100 = 10,000 messages per minute

❓ With a skip limit of 5000, how can the scrapping be improved?

Known Bugs

  • After complete archive of rooms, they're marked isArchived as true.
    Hence, the next time the app is run, it removes the archived rooms from the list.
    The expected behavior is that it should simply ignore the archived rooms and only consider the non archived rooms.
    This results in archive pattern download behaving weird i.e. since it cannot verify if a particular room was archived, it redownloads the room and appends it to the end. (resulting in duplication)
  • HTTP Status 429 rate limiter error. Requests are not being rate limited even if using rate-limiter-api


  • Add cli processor modules