Getting Started
What Astervis is and how to get it running on your server.
Astervis is an on-premise call-center analytics and AI platform for Asterisk and FreePBX. It installs on your own PBX server, replicates every Call Detail Record (CDR) and call recording into a local time-series database, and layers analytics dashboards plus an optional AI pipeline — transcription, QA scoring, sentiment and category analysis, semantic search — on top of the call data you already collect.
The product runs entirely on your infrastructure. Your CDRs and recordings never leave the server, and there is no Astervis cloud to deploy to.
On-premise, always. Astervis is installed on the customer's PBX server and only there. The single exception is AI inference: when you turn on AI analysis, the audio of a finished call is sent to Google Gemini (on Vertex AI) for transcription and scoring. The analytics platform works fully without AI, so if you can't send audio out, you simply don't enable it — see AI & Privacy.
Who it's for
Astervis is built for the people who run a phone team and need to see what's actually happening on the lines.
- Call-center managers get the high-level picture: call volume by hour and weekday, missed and lost calls, sales pulled from your CRM, and operator leaderboards — without exporting spreadsheets out of the PBX.
- Supervisors and QA leads get per-operator performance, average wait time, and — when AI is on — a 0–100 quality score on every recorded call, measured against rules you write yourself.
- Telephony admins get a single installer that detects the PBX, wires up CDR replication, and stands up the whole stack behind one HTTPS port. After install, it mostly runs itself.
If you already have Asterisk or FreePBX answering calls, Astervis turns that raw call data into dashboards your team can act on.
What you get
Analytics dashboards
Call volume, missed/lost calls, operator performance, wait time, trunk load, and sales from CRM — role-aware for admins, supervisors, and operators.
AI call analysis
Every recorded call transcribed, diarized, summarized, and tagged with sentiment, category, topics, and key quotes.
Configurable QA
Score calls 0–100 against your own weighted rubric. Start from 8 industry templates and edit in a visual editor.
CRM & messaging
Two-way Bitrix24 sync, amoCRM OAuth marketplace integration, and omnichannel Open Lines over Telegram and WhatsApp.
Phone blacklist
Block unwanted callers at the PBX with AMI sync to astdb and an ARI hangup fallback. CSV import/export included.
AI analysis and QA scoring are gated by a separate AI license tier. The analytics dashboards, integrations, and blacklist work on every plan, AI or not.
Install in one command
Astervis ships as a single Go installer. The bootstrap script below detects your
OS and architecture, downloads the astervis-installer binary to
/usr/local/bin, and prints the next step. Run it as root on your PBX server:
curl -fsSL https://api.astervis.io/api/releases/install.sh | bashTo download the binary and chain straight into the install wizard, pass your
license key (it starts with AST-, e.g. AST-XXXX-XXXX-XXXX-XXXX...):
curl -fsSL https://api.astervis.io/api/releases/install.sh | bash -s -- --license YOUR_LICENSE_KEYThe wizard walks through four steps — system check, license validation, configuration (admin login, domain/SSL, MySQL), and installation — then prints your access URL. The product is served on a single externally-exposed port:
https://<your-domain-or-IP>:8443Astervis needs a running PBX, at least 2 GB RAM, 1 CPU core, and 15 GB free disk on the Docker data partition, plus outbound internet to pull images. The installer checks all of this up front and installs Docker automatically if it's missing. See Installation for the full requirements, supported platforms, and SSL options.
Under the hood, Astervis runs as a set of Docker containers behind an nginx
reverse proxy. Port 8443 is the only thing exposed to the network — TimescaleDB,
Redis, the backend, the admin UI, and the litcdc change-data-capture service all
bind to 127.0.0.1.
Next steps
Last updated on