Files
cti/p-line/docs/markdown/LOCAL-ASTERISK-CLI-API.md
2026-10-14 23:05:02 +03:30

1.7 KiB

Local Asterisk CLI API

Trusted local programs can submit one Asterisk CLI command per request:

POST http://127.0.0.1:8080/local-asterisk/command
Content-Type: application/json
X-Local-Asterisk-Token: <value of pline.local-asterisk-api-token>

{"command":"core show version"}

Example from a trusted local shell:

TOKEN="$(sed -n 's/^pline\.local-asterisk-api-token=//p' /opt/cti/application.properties)"
curl --fail-with-body -sS http://127.0.0.1:8080/local-asterisk/command \
  -H "X-Local-Asterisk-Token: $TOKEN" \
  -H 'Content-Type: application/json' \
  --data '{"command":"core show version"}'

On installed systems, the token is generated once and stored as pline.local-asterisk-api-token in /opt/cti/application.properties. It remains stable through installer upgrades. Only trusted local processes should be given this value; the API accepts Asterisk CLI commands with the same authority as the pline AMI account.

The service accepts requests only from a direct loopback connection. Requests with proxy-forwarding headers are rejected, so local programs should connect directly to port 8080 rather than through Nginx. Commands are limited to one line and 4096 characters. Responses include the command output and whether AMI or the local CLI handled the request.

CLI commands use a bounded AMI command pool of four connections. If all command slots are busy, the request does not wait in an application queue and may use a bounded local CLI fallback. At most two local CLI fallback processes run at a time; excess requests fail promptly. If an AMI action has been sent but its response times out, the service does not retry the command through the CLI, because its outcome may be uncertain.