CLI
ANALYTICS.

Built for CLI Builders.

Understand package adoption and how your CLI is actually used.

Local-first Open source Self-hostable

See your product from ecosystem to runtime.

Analytics are measurable traces your product leaves behind: in a registry or on a machine. Secchi collects the two families and reads them together.

■ Ecosystem

Packages · Registries · GitHub · Dependencies

adoption · health · releases · security

■ CLI runtime

CLI

usage · versions · failures · performance

SECCHI

Illustrative

■ Package analytics

Understand your ecosystem.

Registry metadata, downloads, releases, dependencies, security and GitHub activity in one view. Across PyPI, npm, crates.io, Homebrew, Go, and CRAN. In the terminal, in reports, or over MCP for coding agents.

PyPI
npm
crates.io
Homebrew
Go
CRAN

$ uvx --from secchi secchi dashboard duckdb

$ uv tool install secchi

$ pip install secchi

Read docs

~/work

Health score 0–100, graded A–F
Adoption trend across 30d / 90d / 1y
Security advisories surfaced automatically

■ CLI usage analytics

Understand how your product is actually used.

Instrument your CLI to understand feature usage, versions, failures and performance without collecting command values or application data.

It only ever records command and flag names, never the values you typed, never error messages. Anything that looks like a secret, an API key, a token, an email, is redacted before it's ever written to disk.

$ cargo install secchi-cli-analytics

let analytics = Analytics::builder("myctl")
    .cli_version(env!("CARGO_PKG_VERSION"))
    .allow_values(["format"])
    .start();

let matches = cli::build().get_matches();
let outcome = run(&matches);

analytics.record_exit(&matches, exit_code, error_class);import click
from secchi_analytics.click import instrument

@click.group()
def cli():
    ...

instrument(cli, app_name="myctl")

Read docs

~/work

Local-first: useful with zero network, zero account
Redaction runs even on values you explicitly allow through
Off with one switch: SECCHI_ANALYTICS=0
Delete everything: secchi analytics purge

A local TUI dashboard for this data ships next.

From analytics to answers.

The same analytics, wherever the decision happens: in your terminal, in a report, on a pull request, inside a coding agent, or as JSON for everything else.

$ uvx --from secchi secchi dashboard duckdb

The local TUI. Health, adoption, versions, and security in one screen. Click to enlarge.

$ secchi report duckdb --format md

# duckdb
Registry: pypi

## Overview
| Analytics            | Value                          |
| -------------------- | ------------------------------ |
| Health score         | 89 / 100 (B)                   |
| Latest version       | 1.5.5                          |
| Downloads (30d)      | 54,469,976                     |
| Adoption change      | +8.3% vs previous 30 days      |
| GitHub stars         | 174                            |

## Health breakdown
| Category       | Score   |
| -------------- | ------- |
| Maintenance    | 20 / 20 |
| Community      |  8 / 15 |
| Documentation  | 11 / 15 |
| Releases       | 15 / 15 |
| Security       | 20 / 20 |
| Testing        | 15 / 15 |

Markdown, HTML, or JSON. Drop it in a PR, a wiki, or a stakeholder deck.

# .github/workflows/secchi.yml
on: [pull_request]

jobs:
  secchi:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v6
      - uses: kannandreams/secchi-action@v1
        with:
          package: pypi:duckdb

Writes the report to the job summary and keeps one PR comment updated across re-runs.

// claude_desktop_config.json, .cursor/mcp.json, ...
{
  "mcpServers": {
    "secchi": { "command": "secchi-mcp" }
  }
}

// tools
inspect_package    health, adoption, releases, dependencies, repo analytics
search_packages    ranked matches across supported ecosystems
inspect_project    a configured project, sources combined
check_package      health and repository CI policies
compare_packages   rank choices with evidence and confidence

A read-only MCP server for coding agents. Same analytics, inside the loop.

$ secchi report duckdb --format json --output -
{
  "schema": "secchi.package-intelligence",
  "schema_version": 1,
  "package": "duckdb",
  "registry": "pypi",
  "package_info": {
    "latest_version": "1.5.5",
    "download_counts": { "today": 1394942, "week": 13604705, "month": 54469976 },
    "github_stats": { "stars": 174, "forks": 102, "open_issues": 53 }
  }
}

Schema-versioned export for automation and agent workflows.

Analytics coverage

CLI analytics, from ecosystem to execution.

■ Ecosystem analytics

Package adoption
Downloads
Dependencies
Releases
GitHub activity
Security

■ CLI usage analytics

CLI usage
Version adoption
Errors
Performance
Human vs. agent

Built in the open, from the first commit.

Run Secchi locally, explore the source, or help shape what comes next.

Get started View on GitHub

CLIANALYTICS.