Questions, answered plainly

A speech API. You can generate speech with 12 live voices, use your own voice, transcribe speech, build voice agents (beta) and play a checked English phrase library. New industry voice packs are coming for every industry.

Developers, software companies, agencies and businesses that need affordable speech in an app, kiosk or phone flow, especially for repeated operational messages.

List prices: catalog voices $0.70, your own voice $6 and styled voice $17 per million characters; speech to text $0.0015 per minute; agents $10 per 1,000 turns. Industry packs are coming and will be priced when published. Every account gets 100,000 free characters, 60 free minutes of transcription and 200 free agent turns. These are list prices.

A set of checked, ready-generated recordings for one industry that you can download and play offline, with a licence. Industry packs (47 live across 23 industries today, more being built for every industry); none are listed on this site until they are ready.

Not through this site yet. Tell us which industry you need first at support@eyeguide.ai.

Live generation covers English, Spanish, French, Hindi, Japanese, Portuguese, Chinese with Kokoro voices; the library is English-first. Non-English library results are labelled until a native speaker approves them. Custom voices support English, with French and German in beta.

Catalog speech returns an existing checked recording instantly. Live generation synthesises any other text, taking seconds for a new sentence. Both are billed per character.

No. Library and repeat requests through the API are billed by character each time. Downloaded pack audio has no per-play cost.

Library audio is public. Custom-voice audio is private to your account, stored only after a sentence is requested twice and deleted after 75 days without use. Live catalog sentences are stored to speed repeats. Do not send sensitive personal information.

A static recording cannot. Use a library line that contains the value, generate the sentence live, or show the variable part on screen.

Speakvora makes no HIPAA, BAA, SOC 2 or clinical-review claim. Do not send patient-identifiable text.

There are single-file JavaScript and Python helper clients at /sdk/speakvora.js and /sdk/speakvora.py. There is no published npm or PyPI package.

Not directly with your API key. Call the API from your server. Browser tokens exist for voice agents only.

Yes, through an MCP server at /api/mcp (tools: speak, list_voices, search_phrases) and an llms.txt file.

Yes: create an account in the developer portal and get an API key.