Every listing call answered from your data
Dograh is the self-hostable voice agent platform with an open-source workflow builder. Answer listing questions from your own feed and book viewings from the live calendar, on servers your agency runs.
Three calls your agents stop chasing
Answer from the listing record
Pre-Call Data Fetchloads the listing from the number dialled
The agent reads price, bedrooms and open-house times from your feed and nothing else. A missing field goes to the listing agent.
→ Give each listing its own tracking number
The same four questions for every buyer
Variable extractionrecords each answer as the buyer said it
Budget range, area, timeline and financing status go to your CRM, which does any scoring.
→ The agent never tells a buyer they do not qualify
Book only slots the calendar returns
HTTP API toolreads and writes your showing calendar mid-call
The agent offers two open times inside the seller's showing window and reads the booking back before writing it.
→ Send workflow_run_id so a repeat request books once
Fair housing rules written into the script
- One script for every callerEvery buyer hears the same questions in the same words and is offered the same open slots.
- Five questions go to a personNegotiation, disclosures, schools, neighborhood make-up and area safety get one fixed reply, then a human agent.
- No area suggestionsThe agent records the areas a buyer names and never describes who lives there.
- Consent before outbound callsA new web lead gets a call only when the form's consent names your agency and the number.
- Offers go to a licensed agentPrice talk transfers on Twilio after the agent writes the lead to your CRM.
Self-host and buyer data stays with the agency
Recordings stay on your servers
Self-host Dograh and recordings and transcripts sit with the agency. Caller audio stays inside your infrastructure only when the models run there too.
Your CRM stays in charge
No built-in CRM integration to outgrow. HTTP API tools, MCP tools and webhooks read and write the system you already run.
Spanish and English on one line
Multilingual auto-detect follows the caller's language, and a short greeting in both languages plays first.
No per-minute platform fee
Self-host the BSD-2 code and pay only for servers, phone minutes and the model keys you choose. Dograh Cloud has a 1¢-a-minute platform fee.
What changes when the platform is yours
| What you get | Dograh | hosted voice AI platforms |
|---|---|---|
| Open-source code you can read and change (BSD-2) | ✓ | ✕ |
| Self-host on your own servers or on-prem without an enterprise contract | ✓ | ✕ |
| Run the whole stack, platform and models, on your own hardware | ✓ | ✕ |
| No per-minute platform fee when self-hosted | ✓ | ✕ |
Every agency call, one deep dive each
Does the agent answer questions about schools or neighborhoods?+×
No. Those get the same fixed reply for every caller and go to a person, along with negotiation and disclosure questions.
Can the agent call new web leads?+×
Only when the lead form's consent names your agency and the number. The FCC treats AI voices as artificial voices under the TCPA.
Which CRM does it work with?+×
Any CRM with an API. Dograh has no built-in CRM integrations, so HTTP API tools, MCP tools and webhooks connect it.
Does call data stay on our servers?+×
Only when you self-host Dograh and run the speech and language models there too. Hosted model keys send audio to that provider.
Can it take calls in Spanish?+×
Yes. Set speech-to-text to Multilingual (Auto-detect), add Spanish and English as Language Hints and write the greeting in both languages.
What does it cost?+×
Self-hosted Dograh has no per-minute platform fee; you pay for servers, phone minutes and model keys. Dograh Cloud charges 1¢ a minute.
Start with one listing line.
Self-host in one command, or start on Dograh Cloud.
$ curl … start_docker.sh && ./start_docker.sh