What did Hlido score Cicada?

Cicada scored 76/100 (STEADY) in Hlido's independent, hands-on review.

How much does Cicada cost?

Hlido recorded the pricing model as: Open source. Derived from evidence Hlido holds, not vendor-supplied; verify current prices on the vendor's pricing page.

Does any vendor pay Hlido for placement?

No. Hlido takes no money from the agents it rates — scoring weights stay private and the evidence behind every verdict is public.

Cicada

Name: Cicada review
Item: Cicada
Rating: 76
Author: Hlido Editor

Coding · tested 2026-08-19 · re-test due 2026-11-19 · by the Hlido desk, not the vendor

In short: AST-powered code intelligence for Elixir (and beta Python) delivered over MCP, with a refreshingly transparent — if vendor-run — benchmark instead of vibes.

Visit Cicada → scorecard.json

Quick answer

Cicada scores 76/100 (STEADY) on Hlido’s independent, hands-on test (reviewed 2026-08-19). STEADY (76) for an MIT-licensed, local, MCP-native code-intelligence tool with a real design (AST symbol access, call graph, semantic search, git attribution) and — unusually — a dated, methodology-published benchmark ra Pricing: Open source (free entry point documented).

Hlido Editor · 2026-08-19

Cicada (cicada-mcp, MIT) gives an AI assistant structured, symbol-level access to an Elixir codebase — modules, functions, signatures, typespecs, docs with precise locations — plus a full call graph and impact analysis, semantic search by concept rather than exact name, and git-integrated attribution (who wrote code, review context, co-change patterns). Python support is in beta. It runs local and private with zero telemetry and installs cleanly via uv (uv tool install cicada-mcp), working with Claude, Cursor, Gemini, VS Code, OpenCode and Codex. What lifts Cicada above the crowded 'code context for agents' field is its willingness to show numbers: a dated November 2025 benchmark against a standard agent on two real repos (4kLoC litmus, 12kLoC oban) reports 20–69% fewer tokens, 27–48% less wall time, and higher accuracy scores, with a linked full report and a published JSON. Hlido treats these as vendor-run and self-judged (scored by an LLM analysing the two end reports), so they are credible-but-not-independent — but publishing the methodology and raw data is materially more honest than the unbacked superiority claims common in this category. The testimonials are an interesting twist: they are attributed to AI tools (Gemini, Codex, Claude Code, Cursor) rather than humans, fitting the agent-as-user framing but not independently verifiable. Reviewed from the marketing and docs surface only; the indexing and query quality were not exercised.

Why STEADY

STEADY (76) for an MIT-licensed, local, MCP-native code-intelligence tool with a real design (AST symbol access, call graph, semantic search, git attribution) and — unusually — a dated, methodology-published benchmark rather than bare claims, discounted because that benchmark is vendor-run and LLM-judged, Python is still beta, the Elixir focus is niche, and the review is surface-only. Low-medium confidence.

What we saw

4 screenshots captured by the Hlido engine during the reviewed run (run-d3ad6e30c5229115-cicada-mcp-vercel-app). Our own captures — not vendor marketing material.

Cicada — run screenshot 1 (home.png) — home.png

Cicada — run screenshot 2 (page_.png) — page_.png

Cicada — run screenshot 3 (page_.png) — page_.png

Cicada — run screenshot 4 (page_.png) — page_.png

What it does well

Structured, symbol-level code access over MCP — modules, functions, signatures, typespecs and docs with precise locations, plus a full call graph and impact analysis
Semantic search by concept (finds verify_credentials when you search 'authentication') and git-integrated attribution for design context
Transparent benchmarking: a dated Nov-2025 test on two real repos with a linked full report and published JSON — method shown, not just a headline
Local and private with zero telemetry; clean uv install; works with Claude, Cursor, Gemini, VS Code, OpenCode and Codex
MIT-licensed and open source

What it fails at

The benchmark is vendor-run and LLM-judged (Sonnet analysing the two end reports) — credible but not independent
Elixir-first with Python only in beta — narrow language coverage versus general-purpose code-graph tools
Testimonials are AI-authored and unverifiable; there is little independent human social proof on the surface
Surface-only review — Hlido did not index a codebase, so query accuracy and completeness are unverified

Best for

Elixir teams who want their AI assistant to navigate, understand impact, and refactor safely without token-burning greps
Developers who value a local, telemetry-free tool with a published benchmark methodology
Python users comfortable adopting a beta feature

Not recommended for

Teams needing mature multi-language coverage today (Python is beta; most languages unsupported)
Anyone requiring independent, third-party benchmark verification
Non-developers — this is an MCP tool wired into a coding assistant

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 2026-08-19.

Related agents

Agent relevance

CLI MCP Behavioral-testable

Agentic-Commerce Readiness 57/100 · INTEGRABLE

Independent readiness for agent delegation & transaction. How it’s scored · check live

Install with `uv tool install cicada-mcp`; it runs as a local MCP server that Claude, Cursor, Gemini, VS Code, OpenCode or Codex query for symbol-level code intelligence, call-graph impact and semantic search.

Agent-friendly score: 9/10

Evidence

Structured, symbol-level code access over MCP — modules, functions, signatures, typespecs and docs with precise locations, plus a full call graph and impact analysis — source (2026-08-19) verified
Semantic search by concept (finds verify_credentials when you search 'authentication') and git-integrated attribution for design context — source (2026-08-19) verified
Transparent benchmarking: a dated Nov-2025 test on two real repos with a linked full report and published JSON — method shown, not just a headline — source (2026-08-19) verified
Local and private with zero telemetry; clean uv install; works with Claude, Cursor, Gemini, VS Code, OpenCode and Codex — source (2026-08-19) verified
Hands-on runtime behaviour (executing the tool / a live task) — source (2026-08-19)

scorecard.json · transparency passport · registry · methodology

More: compare agents · best of · developer tools · incident registry

Verdict by Hlido Editor, our automated editorial system · Method: public-surface-tier-2+editorial-narrative-v2 · Methodology version 2026.05 · Next review due 2026-11-19

How this page was produced. The scores, claim verdicts and evidence come from automated hands-on testing of the product’s public surface. The written analysis is drafted by an AI system, and pages publish without a person reviewing each one. Hlido publishes this record and answers for it — tell us if anything here is wrong and we will correct it.

Embed this trust badge

Live, always-current independent score — free to embed on your site or README. No vendor pays for placement.

Markdown

[![Hlido trust score](https://hlido.eu/badge/wende-cicada.svg)](https://hlido.eu/check/?agent=wende-cicada)

HTML

<a href="https://hlido.eu/check/?agent=wende-cicada"><img src="https://hlido.eu/badge/wende-cicada.svg" alt="Hlido trust score"></a>