Skip to main content
Glama

Footfall

Real usage, not hype. Validated adoption signals for AI developer tools.

CI License Python Code style: ruff Security: Bandit Coverage

Footfall tracks how much real use named AI developer tools get, compared with how much attention they get. Tools means agent frameworks, SDKs, MCP servers and models — each tracked as one entity across GitHub, npm, PyPI and Hugging Face.

It stores immutable daily snapshots, filters unusual star bursts without per-person tracking, classifies each tool's lifecycle phase, and publishes a cryptographically hashed backtest of its own past calls.


Deliverables & Architecture

Deliverable

Location

Description

Remote MCP Server

src/footfall/server/

FastMCP streamable server (footfall_compare, footfall_trend, footfall_find_rising) with DNS rebinding security and byte capping.

Agent Skill

skills/footfall/SKILL.md

Teaches coding agents (Claude Code, Cursor, Windsurf) how to query the tools with provenance caveats.

Static Dashboard

site/

Zero-tracker dashboard (< 11 KB gzipped, strict CSP default-src 'self', WCAG AA contrast).

Syndicate Alerts

site/feeds/

Automated RSS 2.0, Atom 1.0, and JSON Feed 1.1 notifications on lifecycle phase transitions.

Stack Health Analyzer

src/footfall/stack/

CLI utility analyzing pyproject.toml, requirements.txt, and package.json against AI ecosystem momentum.

Adoption Depth Engine

src/footfall/dependents/

Downstream dependent packages analyzer via packages.ecosyste.ms measuring true architectural commitment.

Directory Submissions

submissions/

Complete application packs for Claude Connectors, ChatGPT Plugins, and the MCP Community Registry.

Open Dataset

dataset/

Parquet snapshots and data dictionary published under CC BY 4.0.


Related MCP server: Dev AI Ecosystem MCP

MCP Server Tools

Tool

Parameters

Description

footfall_compare

entities: list[str] (2–6)

Side-by-side comparison of adoption $z$-score, attention $z$-score, divergence ($D$), and lifecycle phase.

footfall_trend

entity: str, weeks: int (4–52)

Historical weekly timeseries for a single tool with velocity rates and anomaly flags.

footfall_find_rising

phase: str, category?: str, limit?: int

Filter tools currently breaking out or accelerating.


CLI Reference

Footfall provides a unified CLI installed via uv:

# 1. Verify cryptographic hash chain of historical calls and backtest
uv run footfall reproduce --verify

# 2. Analyze developer dependencies in your project
uv run footfall stack pyproject.toml
uv run footfall stack package.json --json-out

# 3. Check ecosystem adoption depth (dependent packages vs download traffic)
uv run footfall dependents --source pypi --package langchain-core
uv run footfall dependents --source npm --package @modelcontextprotocol/sdk

# 4. Generate syndicate feeds and dispatch HMAC-signed webhooks
uv run footfall alerts --ledger data/ledger/calls.jsonl --feed-dir site/feeds/

# 5. Export static JSON bundles and shields.io badge endpoints
uv run footfall export --output-dir exports/v1

# 6. Launch the production MCP server (SSE or stdio transport)
uv run footfall server --transport sse --port 8000

Development & Verification

# Install frozen dependencies
uv sync --frozen

# Run full test suite (118 unit, contract, and site tests)
uv run pytest tests/ -v

# Run Ruff linter and formatter
uv run ruff check . && uv run ruff format --check .

# Run static type checking in strict mode
uv run mypy src

# Reproduce backtest from raw data
make reproduce

Status & Roadmap

Phase

Description

Status

0 — Validate

Hypothesis H1/H2 walk-forward evaluation, pre-registration SHA-256 seal, rate-limit verifications.

✅ Passed (Gate G1 Approved)

1 — Data Product

Idempotent collectors, raw store, feature engine, anomaly filter, call ledger, backtest reproduction.

✅ Complete & Hash-Chained

2 — Serving Layer

FastMCP server, container deployment (OIDC), static dashboard, skill prompt suite, security audit.

✅ Passed (Gate G3 Approved)

3 — Launch & Distribution

Claude directory pack, ChatGPT pack, MCP registry entry, launch assets, pilot user guide.

✅ Complete (Gate G4 Approved)

4 — Backlog

Stack analyzer, alert syndication feeds, dependents depth engine, learned classifier benchmark.

✅ Implemented & Verified


Security & Ethics Policy

  • Zero autonomous inspection of secrets: Footfall never stores, inspects, or logs credentials or .env files.

  • Aggregates only: Zero per-person data, GitHub usernames, or individual accounts are stored or logged.

  • Strict Allow-Listing: All API and CLI inputs are validated against strict regex allow-lists (^[A-Za-z0-9@._/\-]{1,100}$).

  • Cryptographic Provenance: Backtest reports and call ledgers are immutable and cryptographically verified via SHA-256 hash chains.

  • Tone Standard: Repositories with inorganic growth are termed unusual, never defamatory.


License

Statistical information derived from public software repositories and package registries. Not investment, purchasing, or software architecture advice.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • F
      license
      Not graded
      quality
      C
      maintenance
      Provides AI agents with 6 tools for searching Hugging Face models, GitHub trending repos, analyzing GitHub repositories, fetching dev.to articles, Show HN launches, and Product Hunt daily launches. Built for dev tooling research and AI ecosystem analysis.
      -
    • F
      license
      Not graded
      quality
      B
      maintenance
      Provides protocol-neutral market intelligence for the AI agent economy, with read-only tools to search agents, retrieve details and histories, compare agents, list categories, view category rankings, and access methodology.
      -
    • A
      license
      Not graded
      quality
      B
      maintenance
      Enables AI agents to query a repository's verified semantic model, including capabilities, ownership, contracts, constraints, impact, health, and verification status, without re-deriving project facts. It exposes read-only tools for inspecting, reasoning about, and verifying the provenance-tracked surface.
      2
      Apache 2.0