Speechmatics
Speechmatics
★ #21 in Audio GenerationProduct status: Status not recorded
No lifecycle review is recorded for this listing. Its inclusion in the directory does not confirm current availability; check the vendor before relying on it.
No status check date recorded
Streaming, batch and voice-agent transcription APIs with cloud and contracted enterprise deployment options. No lifecycle review is recorded for this listing. Its inclusion in the directory does not confirm current availability; check the vendor before relying on it.
Product preview
Decision notes
Official-document research
Based on the cited documentation. We did not use the product, purchase a plan, or run quality, speed or reliability tests in this review.
Consider it for
- Developers who can review transcripts, data consent and usage budgets
Choose another approach if
- One advertised hourly rate for every endpoint or unrestricted software/model resale
Pricing and conditions
Documented terms
USD: read Pre-recorded view Melia-1 0.12/audio hour, Enhanced 0.38, Oak-1 Medical 0.15, Standard 0.23; these are not verified Realtime/Linden rates. One credit equals USD1; new accounts get a one-time USD100 grant. Pro uses credits first, then charges prior-month usage on the next month's first day, measured to the second. Model-training discount is optional opt-in; volume discount applies only above 500h/month per mode+model, not all hours. Enterprise/on-premises quotes separate. Taxes, regional totals, add-on rates, credit expiry and refund conditions unconfirmed; no unlimited allowance inferred.
Pricing source read:
Pricing or terms sourceCode, model and service licensing
Service terms allow integrated products and restrict standalone resale and training competing services. Transcript IP is assigned to the customer; that does not clear third-party recordings, voices or all synthesis outputs. You need lawful input rights and consents. Inputs/outputs are not used for model training unless you enable it in the workspace. Opting in grants broad perpetual irrevocable sublicensable data rights under the Training Addendum, with authorized third-party participation; disabling changes future usage and is not a promise to undo prior training. Processing, portal playback and sharing still require storage; exact retention and an actual Enterprise agreement were not verified. SDK, model and service permissions are separate.
Basis and limits of the recorded score
Original 4.5/5 retained; original score worksheet and hands-on evidence missing. Documentation did not re-score.
Overview
Speechmatics documents Realtime streaming, Batch transcription jobs and Agent STT with speaker-labelled turns for voice applications. Model choice controls language, dictionary and feature support. Cloud access and Enterprise on-premises/on-device deployments have separate terms; access to containers is not an open-source/model-weight license. No accent, noise robustness, latency or accuracy benchmark was run.
Sources
The factual claims in this review were read from the pages below. Each date is when an editor last opened the source, not when the product changed.
- Vendor documentation Checked 2026-10-15
- Vendor pricing page Checked 2026-10-15
- Vendor documentation Checked 2026-10-15
- Vendor documentation Checked 2026-10-15
- Vendor documentation Checked 2026-10-15
- Vendor documentation Checked 2026-10-15
Pros / Cons
Pros
- Separate file, live and agent modes
- Default training choice and optional discount are documented
Cons
- Rates and language/features depend on model/mode
- Training discount changes data rights; enterprise deployment is separately contracted
Who it's for
Reviewed meeting/media transcripts
Developer voice-agent turn streams
Bottom line
Compare the exact processing mode and model, and evaluate the training discount as a data-rights decision.
Key Features
Tags
Frequently asked questions
Is Speechmatics currently available?
No lifecycle review is recorded for this listing. Its inclusion in the directory does not confirm current availability; check the vendor before relying on it.
What is Speechmatics?
Speechmatics is a Audio Generation tool developed by Speechmatics. Streaming, batch and voice-agent transcription APIs with cloud and contracted enterprise deployment options.
What costs and terms should I check for Speechmatics?
USD: read Pre-recorded view Melia-1 0.12/audio hour, Enhanced 0.38, Oak-1 Medical 0.15, Standard 0.23; these are not verified Realtime/Linden rates. One credit equals USD1; new accounts get a one-time USD100 grant. Pro uses credits first, then charges prior-month usage on the next month's first day, measured to the second. Model-training discount is optional opt-in; volume discount applies only above 500h/month per mode+model, not all hours. Enterprise/on-premises quotes separate. Taxes, regional totals, add-on rates, credit expiry and refund conditions unconfirmed; no unlimited allowance inferred.
Is Speechmatics open source?
Service terms allow integrated products and restrict standalone resale and training competing services. Transcript IP is assigned to the customer; that does not clear third-party recordings, voices or all synthesis outputs. You need lawful input rights and consents. Inputs/outputs are not used for model training unless you enable it in the workspace. Opting in grants broad perpetual irrevocable sublicensable data rights under the Training Addendum, with authorized third-party participation; disabling changes future usage and is not a promise to undo prior training. Processing, portal playback and sharing still require storage; exact retention and an actual Enterprise agreement were not verified. SDK, model and service permissions are separate.
What are the best alternatives to Speechmatics?
Notable Audio Generation alternatives include ElevenLabs, Suno, Udio. Browse all 51 tools in our Audio Generation category for a full overview.
Was this listing useful?
Basic Info
- Pricing
- Pay-as-you-go
USD: read Pre-recorded view Melia-1 0.12/audio hour, Enhanced 0.38, Oak-1 Medical 0.15, Standard 0.23; these are not verified Realtime/Linden rates. One credit equals USD1; new accounts get a one-time USD100 grant. Pro uses credits first, then charges prior-month usage on the next month's first day, measured to the second. Model-training discount is optional opt-in; volume discount applies only above 500h/month per mode+model, not all hours. Enterprise/on-premises quotes separate. Taxes, regional totals, add-on rates, credit expiry and refund conditions unconfirmed; no unlimited allowance inferred.
- Editorial score
- ★ 4.5 / 5 How we score
- Platforms
- APIOn-premises
- Last verified
- Not recorded
