OSS Radar
Back

Compare

Compare repos side by side

Pick up to 3 tracked repos to see how they stack up on popularity (stars), momentum (stars gained recently), and maintenance (releases and code pushes).

Loading…

This week's head-to-heads

The two fastest-rising repos in each category over the last 7 days, compared on stars gained, total stars, forks, language, age and releases.

AI agents

rea vs deepseek-harness

rea gained 41,513 stars in the last 7 days against 7,237 for deepseek-harness, about 5.7 times as many.

In total, rea has 399k stars and 100k forks; deepseek-harness has 264k stars and 32k forks.

Both are written mainly in TypeScript. rea was created 7 months ago and deepseek-harness 3 months ago.

rea's latest release is rea-agents-6.8.0 (2 days ago); deepseek-harness has no tagged release yet.

See the full comparison table

LLM tooling

dsh-our-free-model vs graphify

dsh-our-free-model gained 13,826 stars in the last 7 days against 5,634 for graphify, about 2.5 times as many.

In total, graphify has 137k stars and 13.1k forks; dsh-our-free-model has 47.6k stars and 1,324 forks.

dsh-our-free-model is written mainly in JavaScript and was created 33 days ago; graphify is mainly Python, created 7 months ago.

dsh-our-free-model's latest release is v2.7.2 (2 days ago); graphify's latest release is v0.9.97 (today).

See the full comparison table

MCP

hyperframes vs xerj

hyperframes gained 5,022 stars in the last 7 days against 4,429 for xerj, a close race.

In total, hyperframes has 74.2k stars and 6,736 forks; xerj has 17.2k stars and 1,260 forks.

hyperframes is written mainly in TypeScript and was created 8 months ago; xerj is mainly Rust, created 4 months ago.

hyperframes's latest release is v0.8.210 (today); xerj's latest release is v1.0.0-rc.100 (today).

See the full comparison table

Coding agents

orca vs magpie

orca gained 7,205 stars in the last 7 days against 5,391 for magpie, a close race.

In total, orca has 108k stars and 6,848 forks; magpie has 24.2k stars and 1,960 forks.

orca is written mainly in TypeScript and was created 8 months ago; magpie is mainly Go, created 34 days ago.

orca's latest release is v1.4.242 (yesterday); magpie's latest release is v0.1.1804 (today).

See the full comparison table

RAG / vector

DeepTutor vs LEANN

DeepTutor gained 4,745 stars in the last 7 days against 486 for LEANN, about 9.8 times as many.

In total, DeepTutor has 46.8k stars and 5,841 forks; LEANN has 13.5k stars and 1,240 forks.

Both are written mainly in Python. DeepTutor was created 10 months ago and LEANN 17 months ago.

DeepTutor's latest release is v1.6.15 (9 days ago); LEANN's latest release is v0.3.8 (29 days ago).

See the full comparison table

Local LLMs

DigUp vs StartLux-Decision

DigUp gained 589 stars in the last 7 days against 477 for StartLux-Decision, a close race.

In total, DigUp has 5,942 stars and 461 forks; StartLux-Decision has 1,598 stars and 156 forks.

DigUp is written mainly in Swift and was created 19 days ago; StartLux-Decision is mainly Python, created 28 days ago.

DigUp's latest release is v0.6.2 (6 days ago); StartLux-Decision's latest release is v1.1.0 (14 days ago).

See the full comparison table

How to compare open-source repos

Stars
GitHub's bookmark button. A high total means many developers noticed the project at some point; it says little about whether it is still maintained.
Stars gained (24 hours, 7 days)
Momentum. OSS Radar ranks repos by these gains, so a young project gaining fast can outrank an older one with far more total stars.
Forks
Copies people made to change the code. Many forks relative to stars usually means people are building on it, not just bookmarking it.
Last code push and latest release
Maintenance signals. A recent push shows someone is still working on it; tagged releases show versions you can pin to.
Hacker News
How many HN stories linked the repo and how many points they drew: a rough measure of discussion outside GitHub.

Stars, forks, pushes and releases come from the public GitHub API and are refreshed hourly; Hacker News figures come from HN's public search API. Sponsors never affect rankings or comparisons.