bin+lib credit

A fast tool for measuring Github repository contributions

9 releases (stable)

1.4.0 Feb 6, 2021
1.3.0 Aug 24, 2020
1.2.0 Jul 22, 2020
1.1.0 Jun 29, 2020
0.1.0 Jun 19, 2020

#1913 in Command line utilities

Download history 1/week @ 2022-11-30 1/week @ 2022-12-07 9/week @ 2022-12-14 3/week @ 2022-12-21 1/week @ 2022-12-28 1/week @ 2023-01-04 1/week @ 2023-01-18 5/week @ 2023-01-25 6/week @ 2023-02-01 26/week @ 2023-02-08 45/week @ 2023-02-15 1/week @ 2023-03-08

55 downloads per month

MIT license



Build AUR version

credit is a fast tool for measuring Github contributions.

Use credit to find out:

  • Who the most productive developers are in a given country.
  • Who has the most Pull Requests merged to a project.
  • Who engages in the most discussion in Issues and PRs.
  • How long it takes maintainers to respond to and solve Issues.
  • How long it takes to get PRs merged.
  • If a library would be a safe long-term (i.e. maintained) dependency.

Table of Contents


Arch Linux

With an AUR-compatible package manager like aura:

sudo aura -A credit-bin


cargo install credit


To use credit, you'll need a Github Personal Access Token with public_repo permissions. See here for an additional example.

💡 Note: credit calls the GraphQL-based Github v4 API, which has a much higher rate limit than the REST-based v3 API. This allows credit to run quickly and work on projects with a long development history.

You can use credit limit to check your current API query allowance.

Repository Analysis

Markdown Output

By default, credit outputs text to stdout that can be piped into a .md file and displayed as you wish:

> credit repo --token=<token> rust-lang/rustfmt

# Project Report for rustfmt

## Issues

2462 issues found, 2189 of which are now closed (88.9%).

- 1899 (77.1%) of these received a response.
- 1553 (63.1%) have an official response from a repo Owner or organization Member.

Response Times (any):
- Median: 10 hours
- Average: 34 days

Response Times (official):
- Median: 13 hours
- Average: 39 days

## Pull Requests

1821 Pull Requests found, 1650 of which are now merged (90.6%).
168 have been closed without merging (9.2%).

- 1505 (82.6%) of these received a response.
- 1379 (75.7%) have an official response from a repo Owner or organization Member.

Response Times (any):
- Median: 8 hours
- Average: 2 days

Response Times (official):
- Median: 12 hours
- Average: 2 days

- Median: 17 hours
- Average: 3 days

## Contributors

Top 10 Commentors (Issues and PRs):
1. nrc: 2772
2. topecongiro: 1526
3. marcusklaas: 718
4. calebcartwright: 461
5. scampi: 331
6. kamalmarhubi: 120
7. rchaser53: 103
8. cassiersg: 100
9. gnzlbg: 79
10. otavio: 63

Top 10 Code Contributors (by merged PRs):
1. topecongiro: 513
2. marcusklaas: 125
3. calebcartwright: 74
4. nrc: 72
5. scampi: 64
6. rchaser53: 57
7. davidalber: 34
8. kamalmarhubi: 31
9. ayazhafiz: 28
10. sinkuu: 24

💡 Tip: You can pass multiple repos at once to the repo command. The results will be aggregated, which can give a good view of contributions across an organization.

JSON Output

You can also output the raw results as --json, which could then be piped to tools like jq or manipulated as you wish:

> credit repo --token=<token> rust-lang/rustfmt --json

Large Projects

By default, credit queries for Issues and Pull Requests at the same time, which is fast and works well for most projects. For very large projects, however, this can make the Github API unhappy.

If you notice credit failing on projects ones with many thousands of Issues and Pull Requests, consider the --serial flag. This will pull Issues first, and then Pull Requests. --serial allows credit to even work on the Rust compiler itself!

> credit repo --token=<token> rust-lang/rust --serial

Developer Rankings

credit users can be used to determine a rough list of the most productive Open Source programmers in a given country. This reports a similar number to the one seen on a "Contribution Calendar", although contributions to private repositories have been subtracted.

> credit users --token=<token> --location=Switzerland

💡 Note: Due to the nature of the query made to Github, the data fetching will take several minutes to complete.

# Top 100 Open Source Contributors in Switzerland

There are currently 18518 Github users in Switzerland.

  1. oleg-nenashev (7331 contributions)
  2. cclauss (6378 contributions)
  3. dpryan79 (5604 contributions)
  4. peterpeterparker (4869 contributions)
  5. ReneNyffenegger (4722 contributions)
  6. eregon (4415 contributions)
  7. jeremytammik (3864 contributions)
  8. liufengyun (3787 contributions)
  9. swissspidy (3775 contributions)
 10. pvizeli (3706 contributions)
... and so on

As with repo, the --json flag can be used to output JSON data instead.


A configuration file can be specified at your XDG_CONFIG_HOME, which by default is $HOME/.config/credit.toml.

# Your Github Access Token. With this set, you need not pass `--token` on the command line.
token = "abc123"


How accurate is this?

The numbers given by credit are not perfect measures of developer productivity nor maintainer responsiveness. Please use its results in good faith.

Response Times: Particularly in the Open Source world, volunteer developers are under no obligation to respond in a time frame that is most convenient for us the users.

Merged PRs: Without human eyes to judge a code contribution, its importance can be difficult to measure. Some PRs are long, but do little. Some PRs are only a single commit, but save the company. credit takes the stance that, over time, with a large enough sample size, general trends of "who's doing the work" will emerge. Expect weird results for one-man projects or projects that otherwise have a long history of pushing directly to master without using PRs.

User Rankings: What is a "Top Developer" anyway? Since it is possible to artificially inflates one's contribution numbers, credit uses the following assumption to filter out false positives:

Users with both high contribution counts and somewhat high follower counts must be working on something of value.

So, at first only the top 1,000 most followed developers are considered. Afterward, other metrics are applied to arrive at a fair list of the Top 100.

Can I see commit counts too?

Yes! Pass --commits to the repo command. Keep in mind that this requires more data from Github, and so will take longer to complete.

Why do the Median and Average values differ?

Given the presence of outliers in a data set, it can sometimes be more accurate to consider the Median and not the Mean.

In the case of maintainer response times, consider a developer who usually responds to all new Issues within 10 minutes. Then he goes on vacation, and misses a few until his return 2 weeks later. His Average would be skewed in this case, but the Median would remain accurate.

credit doesn't attempt to remove outliers, but might in the future.


~460K SLoC