| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| README.md | 2026-05-25 | 2.8 kB | |
| v2.3.5 -- Real-time token savings, visible to humans source code.tar.gz | 2026-05-25 | 5.5 MB | |
| v2.3.5 -- Real-time token savings, visible to humans source code.zip | 2026-05-25 | 5.7 MB | |
| Totals: 3 Items | 11.2 MB | 0 | |
Real-time token savings, visible to humans. The estimated context-savings metric introduced in 2.3.4 was JSON-only. In 2.3.5 it surfaces as a clean boxed panel on the CLI and is verifiable against a real tokenizer in one flag β so when you reach for code-review-graph to review a change, you can immediately see how much of your context window the graph just kept out.

Highlights
- πͺ Token Savings panel on both
code-review-graph detect-changes --briefand the newcode-review-graph update --brief. Per-category breakdown (Functions / Tests / Risk / Other) that sums exactly to the graph response size. - β
--verifyflag cross-checks the displayed numbers against OpenAI'scl100k_basetokenizer. Calibration shows the estimate stays within +0.5% of real GPT-4 tokens in aggregate across 222 mixed-language source files (data indocs/REPRODUCING.md). - π Deterministic eval pipeline β pinned upstream SHAs, full clones with
returncodechecks, fixed Leiden seed. Two contributors running the benchmark recipe on different machines on different days now produce identical numbers. - π― Multi-hop retrieval benchmark + richer embedding text + identifier-aware search boost lift compound-query accuracy from 0.545 β 0.909.
- π¦
code-review-graph embedCLI subcommand for explicit embedding generation. Previously only reachable via MCP.
What the panel looks like
```text ββββββββββββββββββββββββ Token Savings βββββββββββββββββββββββββ β Full context would be: 12,921 tokens β β Graph context used: 762 tokens β β Saved: 12,159 tokens (~94%) β β Breakdown: Functions 244 Β· Tests 191 Β· Risk 244 Β· Other 83 β ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ ```
Add --verify to grow a Verified (tiktoken) row so the numbers are no longer just an estimate.
Reproduction
End-to-end recipe with canonical numbers in docs/REPRODUCING.md (github.com). All 6 test repos pin upstream SHAs, embeddings are deterministic on CPU, Leiden detection is seeded.