Home / GitHub
Git-tracked repo files (direct, or via git-unlimited-storage wrapper)
github.comPaaSstorage/git-tracked-repo-filesJSONcomparereport an errorprovider JSON
- Free allowance
- 100 MiB hard per file ('GitHub blocks files larger than 100 MiB'); repository size is soft — "We recommend repositories remain small, ideally less than 1 GB, and less than 5 GB is strongly recommended".
- Verdict
- fails: quotathe free allowance is too small to matter
- Cost
- free tier
- Identity
- email?placeholder, not evidence; card unknown, phone unknown
- Binding limit
- per-object-capper file ('GitHub blocks files larger than 100 MiB
- Evidence
- reported3 sources; a third party said so
- Freshness
- stale on 2026-11-05checked 2026-10-21 (inherited); stale on 2026-11-05
| Free tier | 100 MiB hard per file ('GitHub blocks files larger than 100 MiB'); repository size is soft — "We recommend repositories remain small, ideally less than 1 GB, and less than 5 GB is strongly recommended". unit GB; metered on unstated |
|---|
| Price | none stated |
|---|
| Limits | - per-object-cap (binding) per file ('GitHub blocks files larger than 100 MiB source
100 MiB hard per file, 50 MiB warning, 5 GB strongly recommended per repo, ~80 write req/min and 500/hr per token. |
|---|
| Identity | email (inferred, not read) not published |
|---|
| Egress | Unmetered via raw.githubusercontent.com (probed 200) and codeload.github.com (probed 200 on a 270,150,054-byte tarball). Free, anonymous, CDN-fronted. |
|---|
| Per-object cap | per file ('GitHub blocks files larger than 100 MiB (104857600 bytes) |
|---|
| Retention | | Clone cost grows with history — every revision of a file is retained forever, so this is the worst surface for churn |
|---|
| Interface | s3, git git |
|---|
| Tooling | git; wrapper bharatcj/git-unlimited-storage (Python + Node, S3-shaped API, jsDelivr serving path, 0 stars but 77 tests passing) and sueszli/github-lfs-bypass (5 stars). |
|---|
| Composes with | 100 MiB per file x unlimited repos. The wrapper's own honest-limits table caps the practical use at 'anything under a few hundred gigabytes' and calls itself the 'Wrong tool' for 'storage heading past a terabyte'. |
|---|
| Reliability | Clone cost grows with history — every revision of a file is retained forever, so this is the worst surface for churn. Repos above ~5 GB may draw a support email asking for remediation. |
|---|
| Parsed numbers | - has-free-tier 1
- per-object-cap 100 MiB
|
|---|
Evidence 3
More fields 1
| dead end | {"failure_class": "quota","why": "100 MiB hard per file ('GitHub blocks files larger than 100 MiB'); repository size is soft — \"We recommend repositories remain small, ideally less than 1 GB, and less than 5 GB is strongly recommended\".","retry_when": null,"first_recorded": "2026-10-21"} |
|---|
Unknown values 2
- per object cap notes: the field may be stated in a source that was not read
- retention notes: the field may be stated in a source that was not read