Voxta Add a service
Browse the service catalog and install a new LLM, voice, transcription module, or integration.
The Add Service page is the catalog of everything Voxta can plug in: LLMs, voices, transcription, vision, search, hardware, and integrations. Reach it from Manage Services → + Add Services.
What you need for a working chat
To actually have a conversation with an AI character you need at least one of each:
- Text Generation — the character's brain. Drives replies.
- Inside Text Gen, the same service handles three things by default:
- Reply — the actual conversation responses.
- Action Inference — picking which action to fire.
- Summarization — collapsing long history into short summaries.
- Inside Text Gen, the same service handles three things by default:
- Text-to-Speech (TTS) — the character's voice.
- Speech-to-Text (STT) — your microphone → text.
You can mix and match. A common power-user setup:
- OpenAI GPT-4 class for Reply (best conversation quality).
- Cheaper / faster model for Action Inference and Summarization (less critical paths).
- ElevenLabs for TTS (best voice).
- Deepgram for STT (fast, accurate).
See Services overview for the full catalog grouped by category.
Browsing the catalog
Each service tile shows:
- Name and short description.
- Cost indicator — free / cheap / moderate / expensive / self-hosted.
- Quality rank — see the ranking system on the Manage Services page.
- Local / cloud badge.
Click any service to see its config form. The form fields vary per service — API key for cloud services, model path for local services, etc.