Three different products
- Hosted platforms (Vapi, Retell AI, Synthflow, and Speko Agents, ours): you configure an agent through their API or dashboard and they run the whole call.
- Open-source frameworks (Pipecat, LiveKit Agents): code you run on your own infrastructure, with control over every pipeline stage and the operational burden that comes with it.
- Model routers (Speko Router, ours): one API over many STT, LLM and TTS vendors; the model layer that sits under either of the above.
The checkable facts
| Platform | Type | Open source | Self-host | Pricing model |
|---|---|---|---|---|
| Vapi | Hosted platform | No | - | Usage-based, per minute |
| Retell AI | Hosted platform | No | - | Usage-based, per minute |
| Synthflow | Hosted platform | No | - | Annual enterprise agreement |
| Pipecat | Open-source framework | Yes | Yes | Free framework; you pay model and infra vendors |
| LiveKit Agents | Open-source framework | Yes | Yes | Free framework; LiveKit Cloud bills usage |
| Speko Agents (our product) | Hosted platform | No | - | Usage-based, per minute |
Cells reflect each vendor's public site and repositories, last checked 2026-11-10; a dash means we could not verify a clear public answer. Rates change often, so we state the pricing model, never a price.
The models underneath
None of these products makes the models, except where a vendor also sells its own. A cascade call runs the same STT, LLM and TTS engines whichever platform hosts it, and those engines set the floor for latency, accuracy and cost. Evaluate the models first: STT, LLM, TTS, whole-stack cost per solve. We have also measured LiveKit's own inference host through a voice-agent lens: LiveKit Inference for voice agents.
How to choose
- Build vs buy. A hosted platform is the fastest route to a working agent; a framework trades integration time for control of every stage. Before deciding, check whether the platform exposes the model choices that set your cost floor: cost per solve across full stacks.
- Latency-critical. End-of-turn detection, first token and first audio come from the models. Evaluate a platform by which of the measured STT, LLM and TTS rows it lets you run.
- Multilingual. Language failures live at the model layer and follow the model onto any platform. Check Multilingual STT, Multilingual TTS and Multilingual LLM before choosing where to host.
For a deeper comparison of the hosted platforms, see speko.com/voice-agent-platforms.
