hscrape

The Evo Headless Scraper CLI. Installing this package puts the hscrape command on your PATH and installs the evo-scraper skill for Claude Code.

Install

npm install -g hscrape

Then point it at your Evo API and authenticate:

hscrape auth login --base-url https://<your-evo-api> --api-key <your-key>
hscrape auth status

Named profiles (prod / staging / local)

Most people work against more than one environment. Give each one a named profile so you never have to remember which key goes with which URL — and never fire a write at production by accident.

Run auth login once per environment with --profile:

# production
hscrape auth login --profile prod \
  --base-url https://<your-prod-api> --api-key <your-prod-key>

# staging
hscrape auth login --profile staging \
  --base-url https://<your-staging-api> --api-key <your-staging-key>

# a local dev server, if you run one
hscrape auth login --profile local \
  --base-url http://localhost:5099 --api-key <your-local-key>

Check what's configured and which profile is active:

hscrape auth status

╭───┬─────────┬───────────────────────────────┬──────────────╮
│   │ Profile │ Base URL                      │ API Key      │
├───┼─────────┼───────────────────────────────┼──────────────┤
│   │ local   │ http://localhost:5099         │ ********81e8 │
│   │ prod    │ https://<your-prod-api>       │ ********9393 │
│ * │ staging │ https://<your-staging-api>    │ ********da41 │
╰───┴─────────┴───────────────────────────────┴──────────────╯
* selected profile: staging (active)

Two ways to choose a profile:

hscrape auth use staging          # switch the active profile (persists)
hscrape host list                 # ...uses staging

hscrape host list --profile prod  # target one profile for a single command

Recommendation: leave staging as the active profile and pass --profile prod explicitly. Every production command then has the word prod visible in it, which makes an accidental production write much harder — especially with commands that replace a whole config, like mcp update.

Resolution order

Highest priority first — an explicit --profile always beats the environment variables:

1. --base-url / --api-key flags
2. --profile <name> (or EVO_PROFILE)
3. EVO_BASE_URL / EVO_API_KEY
4. the active profile in ~/.evo/config.json

hscrape auth status prints what actually resolved, which is the quickest way to settle "why did that hit the wrong server?".

What you get

• hscrape — a self-contained binary for your platform. No .NET runtime required; only the one binary matching your OS/CPU is downloaded.
• The evo-scraper Claude Code skill, copied to ~/.claude/skills/evo-scraper, so Claude Code can drive the scraper end to end (scrape → schema → extract → search → MCP).

Update

npm install -g hscrape@latest
hscrape --version          # confirm it actually moved

The skill re-syncs automatically the next time you run hscrape.

Keep this current. Commands that write a whole config back — mcp update, host update — send a full body built from the CLI's own compiled types. An older CLI doesn't just skip fields it doesn't know about; it clears values that were already set, with no error. If a setting you configured has mysteriously reverted, check hscrape --version against npm view hscrape version first.

Your profiles in ~/.evo/config.json are untouched by upgrades — you won't need to log in again.

Notes

• Requires Node.js ≥ 16.7 to install (used only for the installer/launcher).
• Credentials live in ~/.evo/config.json, written by hscrape auth login. This package ships no keys or URLs — you authenticate with your own key.
• hscrape --help lists every command.
• Set HSCRAPE_BINARY_PATH to run a specific binary instead of the bundled one.