Cozzo
Find yours
Giggy logo

Giggy review: speech-to-text pricing and Cozzo’s file test

Speech-to-Text & Text-to-Speech

Speech-to-text API with a free queued Batch mode, plus paid Fast and Streaming. Cozzo tested Batch and Fast file transcription on Nov 4, 2026; Streaming was untested. Prices are vendor-stated.

Giggy is a speech API with three transcription modes that share one prepaid credit balance: free queued Batch, paid Fast, and paid Streaming, plus text-to-speech on the same account. Here’s what each speech-to-text mode costs on Giggy’s own pages (read Nov 16, 2026), and what Cozzo measured on Nov 4, 2026, when we sent 40 clean LibriSpeech test-clean clips through Batch and Fast. Those were file transcriptions of read audiobook speech, not phone calls, and we did not test Streaming. Giggy’s founder emailed Cozzo on Nov 4, 2026, asking us to consider Giggy for our speech-to-text guide. This page is editorial and unpaid: Cozzo wasn’t paid, and Giggy isn’t ranked or added to that guide’s picks or price table.

Updated

Cozzo file test
Nov 4, 2026, on 40 clean LibriSpeech test-clean clips (16 kHz, 809 words). Batch word error rate 1.61% (basic normalizer). Median turnaround 3.08 s on Batch and 2.80–2.83 s on Fast. Streaming untested. Not phone audio, and not a ranking

I last checked Giggy’s prices against its own pricing page on November 16, 2026.

Best for
Builders who want a free queued file-transcription path, or low-cost Fast and Streaming speech-to-text (vendor-stated)
Starts at
Batch free; Fast $0.04/hr
Last verified
November 16, 2026
Price
Batch is free; Fast $0.04/hr

Last verified against the vendor’s own page · How we verify · Pricing changesSpot a wrong price? support@cozzo.io · Is this your product? Claim this listing

Giggy speech-to-text pricing (2026)

Giggy’s Speech API docs (read Nov 19, 2026) say paid live input now holds credit in five-minute blocks (667 credits, 1,334 with translation) and extends the hold while audio flows. When this page was first published (docs read Nov 16, 2026), a paid live session reserved 96,000 credits, more than a personal free account’s 10,000 monthly credits, so Cozzo didn’t test Streaming. It is still untested.

  • Batch (queued)

    Free

    What you get
    Queued file transcription, 0 credits. Recordings from 0.5 seconds to 25 minutes (1,500 seconds)
    Good to know
    Subject to request and queue limits. During Cozzo’s Nov 4 test, one request was refused with a 429 until a slot freed: Giggy’s error said you can have up to 3 Batch generations waiting at once
    Billing and limits
    Returns HTTP 202 with a generation_uuid; you poll for the transcript. Not a live streaming path
  • Fast

    $0.04/hr

    What you get
    Paid file transcription at 4,000 credits per input audio hour, using Pro, Studio or Scale credits. Recordings from 0.5 seconds to 12 hours
    Good to know
    Starter and Creator packs cost more per credit ($0.0459 and $0.0434 per hour). Paid STT rounds up to one credit per request
    Billing and limits
    Same poll-for-transcript path as Batch. USD rates are before taxes, free credits or Giggy Grants. No subscription
  • Streaming

    $0.08/hr

    What you get
    Live microphone or uploaded-recording streaming at 8,000 credits per input audio hour, using Pro, Studio or Scale credits. Paid API sessions accept up to 12 hours; browser previews are capped at 30 seconds
    Good to know
    Starter and Creator packs cost more per credit ($0.0918 and $0.0869 per hour). Paid API live input holds credit in five-minute blocks (667 credits, 1,334 with translation) and extends the hold while audio flows, per Giggy’s docs (read Nov 19, 2026)
    Billing and limits
    Live audio uses a WebSocket at /v1/audio/transcriptions. Uploaded-recording Streaming uses server-sent events. Cozzo has not tested Streaming
  • Credit packs

    From $3

    What you get
    One-time prepaid purchases, shared across speech-to-text, text-to-speech and voice creation. Pro, Studio and Scale are 100,000 credits per $1: $10 / 1,000,000 (Pro), $50 / 5,000,000 (Studio), or $100 minimum on Scale
    Good to know
    Starter: $3 buys 261,300 credits. Creator: $6 buys 552,600 credits. Those two cost more per credit
    Billing and limits
    Purchased credits expire 1 calendar year after purchase. Personal accounts top up to 10,000 free credits each calendar month (they top up to that amount, rather than adding 10,000 to the balance). Workspaces get no free-credit grants
  • Text-to-speech (same credits)

    Batch free; Fast $5 / 1M chars

    What you get
    Batch TTS is free, subject to request and queue limits. Fast is $5.00 per 1,000,000 input characters and Streaming $10.00, using Pro, Studio or Scale credits
    Good to know
    Starter and Creator cost more per character. TTS counts Unicode code points in the full input, including spaces and punctuation
    Billing and limits
    Each paid Fast or Streaming TTS request has a 10-credit minimum. Cozzo has not tested Giggy’s voices

List prices in USD from giggy.ai/speech-to-text, giggy.ai/pricing and giggy.ai/docs/speech-api, read Nov 16, 2026, about 4:01 PM ET. Fast and Streaming dollar rates assume Pro, Studio or Scale credits, before taxes, free credits or Giggy Grants. Per-request rounding applies. Source 1Source 2Source 3

What a month costs (list prices, speech-to-text only)

  • 10,000 minutes of recorded audio (about 167 hours) on Fast, at the Pro/Studio/Scale rate: 167 × $0.04 ≈ $6.68. The same hours on Streaming: about $13.36. Batch is free, subject to request and queue limits, so the same 167 hours of Batch would be $0.
  • A $10 Pro pack (1,000,000 credits) covers 250 hours of Fast or 125 hours of Streaming, before per-request rounding.
  • The 10,000 free credits a personal account tops up to each month cover 2.5 hours of Fast or 1.25 hours of Streaming (Cozzo’s arithmetic at 4,000 and 8,000 credits an hour).
  • Cozzo’s Nov 4, 2026 file test spent no money: 0 credits on Batch, and about 700 of the account’s free credits on Fast.

Compare more:Speech-to-text APIs for voice agents, including Cozzo’s Giggy file testEvery speech model side by side

Languages
Giggy’s speech-to-text page lists 27 languages, including English, Spanish, French, German, Chinese (Mandarin), Cantonese, Japanese, Korean, Arabic, Hindi and Portuguese. The same page’s hero also says “25+ languages.” You can use automatic language detection or set the language in the API. Read Nov 16, 2026
Audio format
The Speech API docs (read Nov 19, 2026) list uploaded transcriptions as WAV PCM16, mono, 8 kHz or 16 kHz, and live WebSocket audio as LINEAR16 at 8 or 16 kHz or MULAW (G.711 μ-law) at 8 kHz, which Giggy converts to 16 kHz. When this page was first published on Nov 16 the docs listed 16 kHz only, and Cozzo’s Nov 4 test used 16 kHz files. Cozzo has not tested 8 kHz phone audio on Giggy
Free to start
Batch transcription is free, subject to request and queue limits. Personal accounts also receive free credits that top up to 10,000 each calendar month. No subscription is required
Credits
Purchased credits expire 1 calendar year after purchase. Paid STT rounds up to one credit per request, so many short clips can cost more than the hourly rate suggests

Giggy vs Deepgram and AssemblyAI

Speech-to-text list prices from each vendor’s own pricing page, plus Cozzo’s Nov 4 file test of Giggy. Giggy prices were re-read Nov 16, 2026. Deepgram and AssemblyAI figures are the ones already on Cozzo’s pages, last read on those vendors’ pricing pages on Oct 27, 2026 (and re-checked Nov 9, 2026, with no price change).

Giggy

Recorded files
Batch free (queued, 0.5 s–25 min); Fast $0.04/hr (up to 12 hours)
Live / streaming
Streaming $0.08/hr; paid live input holds credit in five-minute blocks. Untested by Cozzo
Free to start
Free Batch, plus personal credits that top up to 10,000 a month
Cozzo file test, 40 clean 16 kHz clips
Batch 1.61% WER (basic normalizer), Nov 4, 2026

Deepgram

Recorded files
Nova-3 $0.0043/min ($0.26/hr)
Live / streaming
Flux English $0.0077/min ($0.46/hr), $0.0065 on Growth, turn detection built in
Free to start
$200 credit, no card
Cozzo file test, 40 clean 16 kHz clips
nova-3 1.85%, earlier Cozzo round on a different date

AssemblyAI

Recorded files
Universal-2 $0.15/hr; Universal-3.5 Pro $0.21/hr
Live / streaming
Universal-Streaming $0.15/hr; Universal-3.6 Pro Realtime $0.45/hr, billed on session time
Free to start
$50 credit, no card
Cozzo file test, 40 clean 16 kHz clips
1.36%, earlier Cozzo round on a different date

The file-test gaps are 1 to 4 word errors out of 809 words, so they don’t rank these three. LibriSpeech is read audiobook speech that is very likely in the training data of Giggy’s underlying model and others, so these figures likely overstate real-world accuracy. Only 40 clips, with no confidence intervals. Don’t compare them with word error rates from Cozzo’s phone-call benchmark. Sources: giggy.ai/speech-to-text, giggy.ai/pricing and giggy.ai/docs/speech-api, read Nov 16, 2026; deepgram.com/pricing and assemblyai.com/pricing, read Oct 27, 2026.

We don’t rank Giggy against Deepgram or AssemblyAI on this file test: the gaps are 1 to 4 word errors out of 809, and the clips are clean read speech, not phone audio. Try Giggy’s free Batch on your own call recordings before you rely on the number.

Key features

  • Free queued Batch transcription for recordings from 0.5 seconds to 25 minutes, subject to request and queue limits (speech-to-text page)
  • Fast $0.04/hr and Streaming $0.08/hr with Pro, Studio or Scale credits; Starter and Creator packs cost more per credit (pricing page)
  • 27 languages listed on the speech-to-text page (the hero also says 25+), with automatic language detection or an explicit language in the API
  • Accepts 8 kHz phone audio: uploaded WAV PCM16 mono at 8 or 16 kHz, and live WebSocket audio as LINEAR16 at 8 or 16 kHz or MULAW (G.711 μ-law) at 8 kHz, converted to 16 kHz (Speech API docs, read Nov 19, 2026). Cozzo has not tested 8 kHz audio on Giggy
  • Cozzo tested Batch and Fast file transcription on Nov 4, 2026; Streaming was untested. Unpaid editorial listing

Sources

Giggy price history

Changes Cozzo spotted on Giggy’s own pricing page since we started logging on Oct 17, 2026.

  1. Newly published

    Speech API now lists 8 kHz input; live credit hold is now five-minute blocks

    Before: Docs listed uploaded audio as WAV PCM16 mono 16 kHz; a paid live session reserved 96,000 credits on admission

    Now: Uploaded WAV PCM16 mono at 8 or 16 kHz; live WebSocket audio as LINEAR16 at 8 or 16 kHz or MULAW at 8 kHz, converted to 16 kHz. Paid live input holds credit in five-minute blocks (667 credits, 1,334 with translation) and extends while audio flows

    Read in Giggy’s Speech API docs on Nov 19, 2026. Prices unchanged. Cozzo has not tested 8 kHz audio or Streaming on Giggy.

    Source
  2. Newly published

    New listing: Giggy

    Now: Batch is free (queued, 0.5 s–25 min, request and queue limits); Fast $0.04 per input audio hour (4,000 credits) and Streaming $0.08 (8,000 credits) with Pro, Studio or Scale credits; Starter and Creator packs cost more per credit; no subscription; purchased credits expire in 1 year

    Added to Speech-to-Text & Text-to-Speech. Cozzo tested Batch and Fast file transcription on Nov 4, 2026; Streaming untested. Unpaid editorial listing; not added to the STT guide picks or price table.

    Source
All pricing changes →

Giggy alternatives

Comparing Giggy with the speech APIs Cozzo already covers? These pages use the same verified pricing data.

Alternatives in Speech-to-Text & Text-to-Speech

All speech-to-text & text-to-speech →

Questions people ask

How much does Giggy speech-to-text cost?

Batch is free, subject to request and queue limits. Fast is $0.04 per input audio hour (4,000 credits) and Streaming is $0.08 (8,000 credits) with Pro, Studio or Scale credits. The $3 Starter and $6 Creator packs cost more per credit. Paid requests round up to one credit. From giggy.ai/speech-to-text and giggy.ai/pricing, read Nov 16, 2026.

Is Giggy speech-to-text free?

Yes, Batch transcription is free for recordings from 0.5 seconds to 25 minutes, subject to request and queue limits. Fast and Streaming use prepaid credits. Personal accounts also receive free credits that top up to 10,000 each calendar month.

Did Cozzo test Giggy?

Yes, on Nov 4, 2026. We sent 40 clean LibriSpeech test-clean clips (16 kHz files, 809 words) through Giggy’s free Batch and paid Fast file APIs. Batch had a 1.61% word error rate with a basic normalizer. Median turnaround was 3.08 seconds on Batch and 2.80–2.83 seconds on Fast. These were file transcriptions of read speech, not phone calls. Streaming is untested: at the time, Giggy’s docs said each paid live session reserved 96,000 credits, more than our free account held. The full write-up is on our speech-to-text guide. Giggy isn’t ranked or added to that guide’s picks.

Does Giggy support real-time transcription?

Giggy’s docs say yes: live speech-to-text runs over a WebSocket at /v1/audio/transcriptions, and paid API sessions accept up to 12 hours. Paid live input holds credit in five-minute blocks (667 credits, 1,334 with translation) and extends the hold while audio flows (docs read Nov 19, 2026). Cozzo has not tested this path.

Which languages can Giggy transcribe?

Giggy’s speech-to-text page lists 27 languages, including English, Spanish, French, German, Chinese (Mandarin), Cantonese, Japanese, Korean, Arabic, Hindi and Portuguese. The same page’s hero also says “25+ languages.” You can use automatic language detection or set the language in the API. Read Nov 16, 2026.

How long can a Giggy recording be?

Batch accepts recordings from 0.5 seconds to 25 minutes. Fast and uploaded-recording Streaming accept up to 12 hours. Paid live Streaming also accepts up to 12 hours; browser previews are capped at 30 seconds. From Giggy’s speech-to-text page and pricing page, read Nov 16, 2026.

Do I need a subscription, and do credits expire?

No subscription is required. Giggy sells one-time prepaid credit packs. Purchased credits expire one calendar year after purchase. Personal free credits top up to 10,000 each calendar month; workspaces don’t get those grants.

Is this an unpaid listing?

Yes. Giggy’s founder emailed Cozzo on Nov 4, 2026, asking us to consider Giggy for our speech-to-text guide. The guide addition and this tool page are editorial and unpaid. Cozzo wasn’t paid, and Giggy isn’t ranked or added to the guide’s picks or price table.