extend-cli

README

extend-cli

[!WARNING] This CLI is under active development. The command surface, flags, and output formats are not yet stable and may change without notice. Pin to a specific tag if you depend on a particular shape of output.

A CLI for Extend.

Install

Install script (macOS, Linux):

curl -fsSL https://extend.ai/install.sh | sh

Homebrew (macOS, Linux):

brew install extend-hq/tap/extend

npm / npx (no install):

npx @extend-ai/cli --help
# or globally
npm install -g @extend-ai/cli

From source (requires Go toolchain):

go install github.com/extend-hq/extend-cli/cmd/extend@latest

Or grab a signed binary from the releases page. The install script verifies the release checksum and installs to a directory already on your PATH. See its options with:

curl -fsSL https://extend.ai/install.sh | sh -s -- --help

Setup

extend setup

The wizard asks how you want to sign in, through your browser (no API key needed) or with an API key, then picks your region, walks you through the rest, and installs the agent skill. See Authentication for how the two credential types differ.

Authentication

Browser login (extend setup, or extend login directly) is best on your own machine: nothing to create or copy, tokens are stored securely and refresh on their own, and the session is scoped to the workspace and environment you approve. extend logout revokes it; extend whoami shows who you're signed in as.

API keys are long-lived and headless, the right choice for scripts, CI, and agents. Save one with extend setup, or set environment variables directly:

export EXTEND_API_KEY=sk_xxx
export EXTEND_REGION=us                        # us | us2 | eu (default: us)
export EXTEND_WORKSPACE_ID=ws_xxx              # for org-scoped keys

When several sources are configured, credentials are used in this order:

1. EXTEND_API_KEY (or EXTEND_<LABEL>_API_KEY under --env <label>)
2. The API key saved by extend setup
3. The stored extend login session

Check what is in effect with extend whoami or extend config. For token storage details and troubleshooting, run extend help auth.

Use with coding agents

The CLI ships a SKILL.md that teaches agent harnesses (Claude Code, Codex, OpenCode, Cursor, Goose, etc.) to use extend correctly without you having to spell out every command.

Install to the cross-client default path:

extend skill install

This writes ~/.agents/skills/extend-cli/SKILL.md, the path Codex, OpenCode, Cursor, and most other harnesses read, and symlinks it into ~/.claude/skills/extend-cli for Claude Code. To scope the skill to one project instead of your whole machine:

extend skill install --target ./.agents/skills/extend-cli/SKILL.md

Re-run extend skill install after upgrading to pick up new commands and flag changes.

Examples

# extract structured data from a local PDF
extend extract invoice.pdf --using ex_abc

# parse to markdown
extend parse contract.pdf > contract.md

# run a workflow async; poll later
RUN=$(extend workflows run doc.pdf --using workflow_abc -o id)
extend workflows runs watch "$RUN"

# filter JSON with jq
extend extract invoice.pdf --using ex_abc --jq '.output.value.invoice_id' -o raw

# batch
extend extract batch invoices/*.pdf --using ex_abc

Inputs can be a local path (auto-uploads), a file_xxx ID, or an https:// URL.

Commands

extract | classify | split  <input> --using <id>
parse <input>
edit <input> --schema schema.json
detect-form <input>
workflows run <input> --using <id>
<action> batch <inputs>... [--files-from list.txt]

extract|classify|split runs   get | list | watch | cancel | delete
parse runs                    get | list | watch | delete
workflows runs                get | list | watch | cancel | delete | update
edit runs                     get | watch | delete
detect-form runs              get | watch
<action> batches              get | watch
files   upload | list | get | delete | download

extractors  | classifiers | splitters | workflows
    list | get | create | update | versions ...

evaluations         list | get | create
evaluations items   list | get | create | update | delete
evaluations runs    get

webhooks endpoints     | subscriptions   list | get | create | update | delete
webhooks verify

Run extend <command> --help for flags.

Output

-o json|yaml|raw|id|table|markdown overrides the per-command default. --jq '<expr>' filters structured payloads before json, yaml, raw, or id formatting. Data goes to stdout, status to stderr. Honors NO_COLOR and CLICOLOR_FORCE.