Sanka CLI

Thin command-line wrapper for Sanka's public CRM and AI API.

The CLI keeps business logic on the server. It handles:

• Developer API token auth and refresh
• local profile and config management
• request construction for CRM, workflow, and AI endpoints
• table or JSON output
• polling for long-running workflow runs

Install

From PyPI:

uv tool install sanka-cli

Bootstrap script from a pinned release tag:

curl -fsSLO https://raw.githubusercontent.com/sankaHQ/sanka-cli/v0.1.3/scripts/install.sh
sh install.sh

From a pinned GitHub release tag:

uv tool install "git+https://github.com/sankaHQ/sanka-cli.git@v0.1.3"

Homebrew support is published through sankaHQ/homebrew-cli. Install with:

brew tap sankaHQ/cli
brew install sankaHQ/cli/sanka

If you previously installed sanka from the old sankaHQ/tap tap (sankaHQ/homebrew-tap), remove that formula and untap it first:

brew uninstall sanka
brew untap sankaHQ/tap
brew tap sankaHQ/cli
brew install sankaHQ/cli/sanka

Authenticate

Create a Developer API Token in Sanka, then save it locally. Login verifies the token against the API and prints the workspace and user it belongs to; a rejected token is not saved. If the API is unreachable, the token is saved with a warning so offline setup still works.

sanka auth login --access-token "<ACCESS_TOKEN>"

Check the active profile:

sanka auth status

Command Areas

CRM records:

sanka companies list
sanka contacts get <contact-id>
sanka deals create --data @deal.json
sanka tickets delete <ticket-id>

Workflow automation:

sanka workflows list
sanka workflows run <workflow-ref>
sanka workflows run <workflow-ref> --wait

Custom Code — write a workflow action in JavaScript or Python, keep it in your own repository, and push versions to Sanka:

sanka code init --slug enrich-company --runtime node
sanka code create
sanka code push --activate -m "add tier logic"

Your repository is the source of truth. Versions are immutable and content-addressed, so pushing unchanged code returns the version that already exists — safe to run on every CI merge. Rolling back is a pointer move, not a redeploy:

sanka code versions enrich-company
sanka code rollback enrich-company          # or: deploy --version 4

pull writes a deployed version back to disk and diff reports whether your working directory matches what is live. diff exits non-zero when they differ, which makes it usable as a CI drift check:

sanka code pull enrich-company --dir ./functions/enrich
sanka code diff --dir ./functions/enrich

Secrets are per-function and never readable back — list shows only a masked suffix:

sanka code secrets set enrich-company CLEARBIT_API_KEY   # prompts, no echo
sanka code secrets list enrich-company

Files are excluded from a push by .sankaignore; .git, node_modules, __pycache__, .venv, build output, and .env are excluded by default.

AI helpers:

sanka ai score company <record-id>
sanka ai score deal <record-id> --score-model-id <score-model-id>
sanka ai enrich company <record-id> --force-refresh
sanka ai enrich company --seed-name "Acme" --seed-url "https://acme.example" --dry-run

Output Modes

The CLI defaults to table output on a TTY and JSON otherwise. Override this per command when needed:

sanka --output json companies list

Environment Overrides

• SANKA_PROFILE
• SANKA_BASE_URL
• SANKA_ACCESS_TOKEN

These override stored profile values without persisting them.

Docs

• Install
• Commands
• Release