scheidydude/codeindex
Coding · tested 2026-09-24 · by the Hlido desk, not the vendor
In short: A temporal code-knowledge-graph tool with blast-radius scoring and a symbol index that materially cuts an agent's token use on codebase tasks — a well-thought-out, dependency-light developer utility.
9 PASS · 1 FAIL of 10 public-surface claims
Quick answer
scheidydude/codeindex scores 82/100 (STEADY) on Hlido’s independent, hands-on test (reviewed 2026-09-24). STEADY reflects an actively maintained, well-documented tool (pip install, CI, releases, recent commits) solving a genuine agent-cost problem with a clean, dependency-light design. Pricing: Open source (free entry point documented).
codeindex builds a persistent SQLite knowledge graph of a repository — dependency graph, per-file blast-radius scores (how many files break if this one changes, including historical as-of queries), a symbol map, and hybrid semantic search — and exposes it ten ways: CLI, Markdown report, an MCP server with 10 tools, a pre-commit hook, and CLAUDE.md injection. The core idea is sound and addresses a real cost centre: instead of an AI assistant scanning every file to locate a symbol, it does an O(1) lookup and opens only the relevant file, which the project claims cuts token usage 60–90% on symbol-location tasks. That claim is plausible in mechanism, though Hlido did not measure it. The engineering choices are pragmatic — no build step, no npm, zero required runtime dependencies because SQLite is stdlib — and it detects 12+ languages. Repo-health signals are strong (documented install via pip, CI, releases, recent commits). Its weaknesses are the honest ones of a young single-author tool: the blast-radius and semantic-search quality will vary by language and codebase, and the value is entirely as a supporting index, not a standalone agent.
Why STEADY
STEADY reflects an actively maintained, well-documented tool (pip install, CI, releases, recent commits) solving a genuine agent-cost problem with a clean, dependency-light design. Not higher because this is a repo-health + public-surface assessment (medium confidence): the headline 60–90% token-saving figure is the project's own, not independently measured here, and analysis quality across 12+ languages is asserted rather than verified.
Public-surface checklist
- PASS Repo reachable (required) — GH API 200 for scheidydude/codeindex
- PASS Readme present (required) — README length 23865
- FAIL License present (required) — NOASSERTION
- PASS Install documented (required) — install/usage section found in README
- PASS Active 12mo (required) — last push 27d ago
- PASS Releases present — latest v0.3.0
- PASS Community traction — 268 stars
- PASS Ci or tests — 1 workflow file(s)
- PASS Recent commit 90d — last push 27d ago
- PASS Agent consumable — MCP server markers in README
What we saw
1 screenshot captured by the Hlido engine during the reviewed run (run-15953b2bff5e8c0c-github-com). Our own captures — not vendor marketing material.
What it does well
- Directly attacks agent token cost: symbol index enables O(1) lookups instead of full-repo scans
- Temporal graph with git-history backfill and historical as-of blast-radius queries — genuinely differentiated
- Zero required runtime dependencies (SQLite stdlib), no build step, no npm
- Ten consumption paths including an MCP server (10 tools), CLI, pre-commit hook, and CLAUDE.md injection
- Strong maintenance signals: pip-installable, CI, tagged releases, recent commits
What it fails at
- The 60–90% token-saving claim is the project's own; Hlido did not benchmark it
- Analysis quality across 12+ languages will vary and isn't independently verified here
- Value is purely as a supporting index — it is not an agent by itself
- Young, apparently single-author project: long-term maintenance risk
- No live demo; assessment is of the repo and documented behaviour
Best for
- Developers using Claude Code / Cursor on large codebases who want to cut token spend on symbol-location tasks
- Teams wanting blast-radius awareness (what breaks if I change this file) surfaced to their AI assistant
- Polyglot repos that need a language-agnostic dependency and symbol index
- Anyone who values a dependency-light, pip-installable tool over a heavier indexing service
Not recommended for
- Teams wanting a turnkey agent rather than a supporting index
- Buyers who need the token-saving and analysis-accuracy claims independently benchmarked before adopting
- Environments that require vendor-backed support guarantees
- Very small repos where full-file scanning is already cheap
Pricing & access
- ModelOpen source
- Free entry pointYes — a free tier or open-source edition is documented
Derived from Hlido-held evidence only (engine checklist + editorial text); quotes are verbatim from the scorecard; not vendor-supplied; re-derived daily. Verify current prices on the vendor's pricing page. Last verified 2026-07-16.
Related agents
Agent relevance
CLI MCP Behavioral-testable
Agentic-Commerce Readiness 53/100 · INTEGRABLE
Independent readiness for agent delegation & transaction. How it’s scored · check live
Ships an MCP server with 10 tools plus CLI, pre-commit hook, and CLAUDE.md injection, so coding agents can query the code graph directly. Designed to be consumed by AI-assisted development workflows.
Agent-friendly score: 8/10
Score over time
The longitudinal record — every point is the score as published on that date. Raw series.
Evidence
scorecard.json · transparency passport · registry · methodology
Embed this trust badge
Markdown
[](https://hlido.eu/check/?agent=scheidydude-codeindex)HTML
<a href="https://hlido.eu/check/?agent=scheidydude-codeindex"><img src="https://hlido.eu/badge/scheidydude-codeindex.svg" alt="Hlido trust score"></a>