Ordalin
Submit a tool

Website-checked tool profile

Dograh

Build and run voice AI agents with an open-source, self-hostable platform.

Sources checked

Visit website ↗
Pricing
The self-hosted open-source platform is free. Managed cloud is pay-as-you-go at a platform fee of 1¢ per minute, with self-serve credits from $5. Committed-volume and on-prem Enterprise plans are priced on request. Official source 1 ↗
Available as
Self-hosting runs with Docker on macOS 10.15 or newer, Windows 10/11 with WSL2, or Linux, on x86_64 or ARM64 CPUs. · Dograh can also be used as a hosted cloud service, with the same MCP and API endpoints at app.dograh.com. Official source 1 ↗ Official source 1 ↗ · Official source 2 ↗
Primary group
Automation & Agents

Also consider

See all 6 similar tools ↓

What it does

Dograh is an open-source platform for building voice AI agents with a drag-and-drop workflow builder. You choose the speech, language-model, voice and telephony providers, then run the stack yourself, on Dograh's managed cloud, or inside your own VPC with Dograh operating it.

Official source 1 ↗ · Official source 2 ↗

Features

  • Swap providers in each step. Each part of the call pipeline (inbound channel, speech-to-text, LLM, text-to-speech and telephony) can use a different provider, or the whole pipeline can switch to a single speech-to-speech model such as Gemini 3.1 Flash Live or GPT Realtime 2. Official source 1 ↗
  • Build agents from a coding assistant. Dograh exposes an MCP server for Claude Code, Codex, Claude Desktop, Cursor and other MCP clients. An assistant can list agents, read their definitions, search the docs and edit agents; edits are saved as a draft until you publish them in the Dograh UI. Official source 1 ↗
  • Python and TypeScript SDKs. Official Python and TypeScript SDKs wrap the REST API and workflow builder, so you can create or edit agents, place outbound calls and inspect runs from your own code. Official source 1 ↗
  • Mix recorded lines with TTS. Agents can play pre-recorded human voice clips and fall back to text-to-speech in the same cloned voice when no recording fits. Official source 1 ↗
  • Self-hosted models and on-prem deployment. Dograh can run on-prem or inside your cloud VPC, with models served inside your perimeter or fully air-gapped. Self-hosted speech models such as Whisper, Voxtral and Kokoro can be plugged in. Official source 1 ↗
  • Telephony and outbound campaigns. The docs cover inbound calling, bulk outbound call campaigns and telephony integrations including Twilio, Vonage, Plivo, Telnyx, WhatsApp and Asterisk ARI. Official source 1 ↗

Pricing & free limits

Pay as you go

1¢ per minute platform feeUsage-based, prepaid credits from $5

Managed cloud with bring-your-own-key or Dograh Models usage pricing. Includes 10 concurrent calls; higher concurrency on request. Official source 1 ↗

Custom volume

Custom pricingCommitted monthly volume

Volume discounts on committed minutes, plus priority support by email, Slack and WhatsApp. Official source 1 ↗

Enterprise

Custom pricingContact sales

On-prem or private-cloud deployment with data residency, SSO/SAML, audit logs, compliance support, a dedicated account manager and custom SLAs. Official source 1 ↗

Billing details

Managed cloud usage is paid from self-serve credits starting at $5, which can be topped up at any time. You can bring your own provider keys or use Dograh Models with usage-based pricing. Official source 1 ↗

Free access & limits

The self-hosted open-source edition is the full platform, not a trial, and the pricing page lists no usage limits for it. Official source 1 ↗

Best for & limitations

Ask a coding agent to build an outbound calling agent for EMI reminders and collections, then add a CRM lookup before each call. Official source 1 ↗

Run voice agents for regulated sectors such as healthtech, insurance or banking on-prem, so calls, recordings and transcripts stay inside your own boundary. Official source 1 ↗

  • Self-hosted LLM swapping is in beta. For self-hosted models, speech-to-text and text-to-speech are production-ready, while swapping the LLM is in beta. Official source 1 ↗
  • Local setup needs 8 GB of RAM. Running Dograh locally requires at least 8 GB RAM (4 GB available to Docker), 10 GB free disk and 2 CPU cores. Official source 1 ↗
  • Concurrency cap on pay-as-you-go. Pay-as-you-go includes 10 concurrent calls; higher concurrency requires contacting Dograh. Official source 1 ↗
Dograh product screenshot
Website previewFixed desktop viewport · 1440 × 900

Similar tools

Published tools that share its categories.

CedricAI teammate in Slack, Telegram or Discord that attends meetings and acts across your tools, with approval before anything sends.
Automation & AgentsFree planChecked Nov 5
Visit
ClickManaged agents API that runs agents with any model, a sandbox, browser and your tools, streaming results to your app.
Automation & AgentsFree planChecked Nov 5
Visit