An AI leasing agent that quotes your PMS
Dograh is the self-hostable voice agent platform with an open-source workflow builder. Your AI leasing assistant quotes live rent and open units, then writes a guest card to your PMS from servers you run.
One prospect call, from hello to guest card
Quote only what the PMS returns
HTTP API toolthe agent asks your PMS mid-call
Open units, rent, fees and specials come back from your PMS mid-call. If nothing fits, the agent tries another size or date.
→ Keep pet and parking rules in the Knowledge Base
Write the guest card as said
Variable extractionrecords each answer as the prospect stated it
Name, move-in date, unit size and pets go to your PMS through a second HTTP API tool. The transcript keeps the exact words.
→ Send workflow_run_id so a retry writes one card
Leave rent talks to a person
Transfer Callthe agent dials your leasing office
Concessions and complaints go to a leasing agent on Twilio after the guest card is written. No summary is passed.
→ If nobody answers in 30 seconds, log a callback request
Reads your PMS, writes back to it
Policies live in the Knowledge Base
The Webhook node retries temporary failures
Call transfer to a person runs on Twilio only
Fair housing rules written into the call
- Same answers for every callerFair housing law generally applies to the agent's words, so every prospect hears the same PMS data and open tour slots.
- One fixed replyQuestions about kids, neighbors, schools or safety get one Transition Speech reply and an offer of a person. Test it on real questions.
- No screening by voiceIncome, credit, rental history and identity checks stay in the written application.
- Accommodation requests to staffAn assistance animal under a no-pets rule is recorded as stated and routed to a person.
- Counsel signs offMany states and cities protect more characteristics, such as source of income. General information, not legal advice.
Prospect data stays with the operator
Recordings on your servers
Self-host Dograh and recordings and transcripts stay with you. Audio stays in-house only when the models run there too.
One agent, every community
Each community's number points at the same agent, and Pre-Call Data Fetch loads that property's name and office hours.
Works through your PMS's own API
No built-in PMS integration to outgrow. HTTP API tools, MCP tools and webhooks do the reading and writing.
No per-minute platform fee
Self-hosted Dograh has no platform fee; servers, model keys and phone minutes still cost. Dograh Cloud charges 1¢ a minute.
Every property call, one deep dive each
What does an AI leasing agent do on a prospect call?+×
It quotes open units, rent, fees and policies from your PMS during the call, logs a guest card and hands the tour booking to your scheduling flow.
Does fair housing law apply to an AI leasing assistant?+×
Generally yes, because it speaks for the property, so every caller gets the same answers and open slots. This is general information, not legal advice; confirm with counsel.
Will the AI leasing agent screen applicants?+×
No. It can read your written criteria the same way to everyone, but income, credit, rental history and identity checks stay in the written application.
Does it connect to our property management system?+×
Through the API your PMS already offers. Dograh has no built-in PMS integration; HTTP API tools read and write during the call.
Can one AI leasing assistant answer for several communities?+×
Yes. Point each community's number at the same agent. Pre-Call Data Fetch receives the dialled number and loads that property's name and hours.
Does call audio stay on our servers?+×
Only when you self-host Dograh and run the speech and language models there too. With hosted model keys, the audio reaches that provider.
How many calls can it take at once?+×
10 by default. Self-hosted, raise it with DEFAULT_ORG_CONCURRENCY_LIMIT; on Dograh Cloud, contact Dograh. A call over the limit is rejected at once, not queued.
Put one community's leasing line on a test number first.
Self-host in one command, or start on Dograh Cloud.
$ curl … start_docker.sh && ./start_docker.sh