radar: repository index
skip to content
Navigate repositories.<br>Spend less context.
radar builds committed maps and exact source routes. Coding agents read fewer files, spend fewer tokens, and finish navigation sooner.
install radar<br>read the evidence ↗
radar query
$ radar query "token verification"
FINAL SOURCE ANCHOR
src/auth.rs#verify_token
28.74 ms fresh query p95<br>10,000 files
40 / 40 top-three<br>source anchors
60.3 % less uncached input<br>verified source packets
57.6 % smaller agent skill<br>368 to 156 tokens
measured advantage
Wins by<br>doing less.
Radar builds the smallest useful index, returns a source pointer instead of a repository tour, and keeps models out of the lookup path.
methods + measured results ↗
same host / 300 Python files ↗
37.1x<br>faster clean build
0.056 s radar map versus 2.069 s code-review-graph build.
Why: compact maps, not a full review graph.
same host / fresh CLI p95 ↗
113.6x<br>faster lookup
2.825 ms radar route versus 321.052 ms code-review-graph search.
Latest query internals: scorer-heavy fresh-query p50 is now 14.65%-25.90% faster; tie-break-heavy route lookup is 79.62% faster.
frozen 40-case ledger / native contracts ↗
15.176x<br>fewer actual tokens
Radar used 919 actual cl100k_base tokens versus TokenSave 13,947; Semble used 86,009, or 93.590x Radar's volume.
Why: exact anchors are smaller than bundled context or semantic snippets.
7 tasks × 3 runs per condition / median sums ↗
63.7%<br>lower reference cost
$0.056409 with radar versus $0.155542 with direct source search; uncached input fell 72.4%.
Why: the agent invokes one exact query instead of exploring files.
agent prompt surfaces / current batch ↗
19<br>accepted token cuts
The generated skill fell 368 → 156 tokens, the navigation contract fell 835 → 462, route-add wrapper keys fell 577 → 152, and a 100-1,000-search daily session projects 159,400-1,594,000 input tokens saved.
Why: Radar budgets the instructions and fallback files agents actually read.
100-1,000 code searches / one session per day ↗
$9.56-$95.64<br>monthly Sonnet 5 input saved
Projection: $0.4402-$4.4020 → $0.1214-$1.2140 per 100-1,000-search daily chat session, or $2.23-$22.32/week.
Why: fewer exploratory source reads become fewer billable input tokens.
Scope matters. The external-tool rows compare different native output contracts on pinned versions; the agent-cost row is a published-rate reference scenario, not a Spark invoice. Broader graph and editing tools can be the better choice when an anchor is not enough.
Fewer files read.<br>Lower agent cost.
Exact routes replace repository tours. Less context goes in, and the source stays in charge.
01<br>Committed maps.<br>Repository structure stays beside the code.
02<br>Verified source.<br>Exact definition packets reject stale answers and whole-file reads.
03<br>Git-aware impact.<br>Callers, tests, dependencies, and co-change history.
04<br>Evidence before complexity.<br>Abstention looked safe on 34 calibration answers, then returned 4 wrong anchors among 9 holdout directs. Radar published it and rolled it back.
05<br>Prompt-budgeted surfaces.<br>Skills, MAPs, route caches, symbol indexes, MCP schemas, and overview rows are measured as agent input, not treated as free text.
Choose your shape.
Different tools, different jobs. This is where radar is the sharper fit.
Open source repository intelligence alternatives<br>projectshapesignalradar boundaryradar is better when
radar ↗reproducedmaps, lexical index, impact, MCP28.74 ms p95 / 40 of 40 top-threeLocal, deterministic, source-verifiable navigation.A source pointer is enough.<br>code-review-graph ↗same-host smokeAST graph, blast radius, PR reviewradar: 37.1x faster build / 113.6x faster fresh p95 lookupBroader review prediction and language coverage.You need verified local routes and a lighter index.<br>TokenSave ↗pinned native-contract rerunsemantic graph, tools, edits9/40 top-one, 14/40 top-three, 13,947 actual tokens; Radar: 15.176x fewerBroader context and agent integrations.You need one bounded source result.<br>Semble ↗pinned native-contract rerunembeddings plus BM25 search12/40 top-one, 15/40 top-three, 18/40 retrieved, 86,009 actual tokens; Radar: 93.590x fewerSemantic chunk retrieval is its strength.You need semantic chunk search.<br>codebase-memory-mcp ↗project-reportedknowledge graph / 15 MCP tools158 languages claimedMuch broader parser and service surface.You need a broad language graph.<br>CodeGraph ↗project-reportedsemantic graph, vectors, MCP32 languages claimedBroader graph, security, and data-flow work.You need data-flow or security analysis.<br>Aider repo map ↗documentedtree-sitter map in an AI loop1k token default map budgetContinuous chat adaptation is its strength.You need chat-adaptive context.<br>Repomix ↗different shapewhole-repository context packno retrieval comparisonBest when a complete portable snapshot is wanted.You need the full repository snapshot.<br>RepoWise ↗project-reportedgraph, Git history, healthunder 30 s / 3,000 files designedDecision and...