Speechify's free plan caps playback at 1.5x with 10 robotic voices and text to speech only. Premium costs $29 a month or $139 a year. The SpeechifyAI API is free up to 500,000 characters a month, then $10, $99 or $499 a month. Dograh, the open-source AI voice agent platform, runs Speechify voices on calls, self-hosted or on Dograh Cloud.
Key Takeaways
- The free plan stops at 1.5x and has no natural voices or AI tools.
- Paying $29 monthly for a year costs $348, against $139 for the yearly plan.
- The API bills per character, from $6 to $10 per million on Simba 3.2.
This post is part of our guide to Speechify alternatives, where Speechify sells a reading app and a separate developer API on two bills. Below, we cover what the free version stops you doing, each paid plan, the API's rates, who each plan suits and voice costs on business calls.
What the free version stops you doing
Speechify's free plan reads text aloud and stops there. It reads at speeds up to 1.5x, with 10 robotic sounding voices and text to speech features only.
Anything faster than 1.5x needs Premium, which goes up to 5x. The voices change too. The free plan has 10 robotic ones, while Premium has more than 1,000 natural voices in over 60 languages.
Everything beyond text to speech is paid as well. Scan & Listen, which reads printed pages through your camera, sits behind Premium. So do AI Summaries & Chats, Voice Typing, AI Podcasts, the Voice AI Assistant and imports from Google Drive, Dropbox and OneDrive.
For a few articles a week, the free plan does the job. Once you want natural voices or listen above 1.5x, it stops fitting.
Premium: monthly or yearly
Premium is one plan with two ways to pay: $29 a month, or $139 billed yearly, which Speechify badges as a 60% saving. By our arithmetic, $139 a year works out at about $11.58 a month. Twelve monthly payments of $29 come to $348. Paying monthly only makes sense for a month or two, to see whether you use it every day.
Speechify offers a free trial of Premium. If you start one, note the date it converts and check the renewal terms before it ends, so you choose the monthly or yearly plan yourself.
Open Source Alternative to Vapi / Retell
Self-hosted voice agent platform — no per-minute fees
dograh-hq/dograh
Star on GitHub
SpeechifyAI API plans and rates
The developer API is a separate product on a separate bill. A Reader subscription does not include API usage, and SpeechifyAI sells five plans, each with a monthly balance of characters.
| Plan | Monthly price | Characters a month (Simba 3.0 or 3.2) | Voice cloning | Concurrent requests |
|---|---|---|---|---|
| Free | $0, no card | 500K | No | 3 |
| Starter | $10 | 1.9M | Yes | 15 |
| Pro | $99 | 13.5M | Yes, no per-plan limit | 30 |
| Scale | $499 | 78M | Yes | 60 |
| Enterprise | Custom | Custom | Custom | 100 |
Simba 3.0 and 3.2 bill $10, $8 and $6 per million characters on Starter, Pro and Scale. Simba 3.5, Speechify's newest model with 48 languages, lists at $20, $16 and $12. Until 21 November 2026 it bills at the same rates as 3.0 and 3.2. After that date, the same balance covers half as many Simba 3.5 characters, so a month on Scale drops from 78M to 39M of them.
Paid plans keep working past their monthly balance through top-ups at the same rates, and top-ups never expire. The monthly balance does not roll over. The Free plan pauses until the next month instead, and it cannot be topped up. Billing counts the characters spoken, so SSML tags are not counted.
The concurrency column matters for anything live. Speechify's API limits are set per account and shared across its synthesis endpoints, and a long stream holds its slot until it finishes. Enterprise adds SSO, SOC 2 Type II and a custom data processing agreement.
Which Speechify plan suits you
Each plan fits a different kind of user. Casual readers can stay on Free. Daily listeners and students should go straight to the yearly Premium plan, and our Speechify review explains who gets the most out of it.
On the API side, Free suits a prototype or a weekend test, with 500,000 characters and three requests at a time. Starter at $10 suits a small app that needs cloning, though 15 concurrent requests cap how many people can hear it speak at once. Pro and Scale are for products with steady volume, where the lower per-character rate and higher concurrency start to count.
Enterprise is for teams whose buyers ask for SSO, SOC 2 Type II or a custom data processing agreement.
Join the Dograh Community
Dograh is an OSS alternative to Vapi. Join our Slack community for queries, releases, best practices & community interactions.
Voice costs on business calls
For a business that wants Speechify's voice on its phone lines, the voice is one line of several on the bill. The call also needs speech-to-text, a language model, phone minutes and a platform to run the agent.
In Dograh, Speechify is a built-in voice provider, set up in the voice settings with your own key. You pick Simba 3.2 or Simba 3.0, so a Speechify voice bills at $10, $8 or $6 per million characters on your own Speechify plan.
| Voice option in Dograh | Per-character voice fee | What you still pay for |
|---|---|---|
| Speechify Simba 3.2 or 3.0, your key | $10 to $6 per 1M characters, on your Speechify plan | The platform, the language model, speech-to-text and phone minutes |
| Kokoro 82M on your own servers | None | The server that runs it, plus the same other lines |
| Pre-recorded clips for fixed lines | None on those lines | Recording them once |
Kokoro is the open model in that table, released under an Apache-2.0 licence, and in Dograh it runs through Local Models (Speaches) with no per-character fee, while the server still costs money. Our guide on how to run Kokoro TTS locally walks through the setup.
Test it on your own script first. Our TTS verbalization benchmark measured how seven open models read tricky text, and kokoro-82m read prices correctly with a score of 0.97 but dates at only 0.10, while the best date score of any model was 0.23.
The platform has its own line too. Self-hosted Dograh has no per-minute platform fee, but servers, model keys and phone minutes still cost money. Dograh Cloud charges a 1¢-a-minute platform fee.
Pre-recorded clips are a second lever on the voice bill. Lines your agent says on most calls play as recordings, with no per-character charge, and the voice handles the rest. Our post on how to cut voice AI costs with pre-recorded audio works through what that saves.
Glossary
- Per-character billing
- Paying for a voice API by the number of characters it speaks. SpeechifyAI counts spoken characters only, so SSML tags are not billed.
- Concurrent requests
- How many voice requests an account can run at the same moment. On SpeechifyAI the limit is per account, from 3 on Free to 100 on Enterprise, and a long stream holds its slot until it ends.
- Top-up
- Extra API balance bought up front on a paid SpeechifyAI plan, at the plan's own rates. Top-ups never expire, while the monthly balance does not roll over.
- Open-weight voice model
- A text-to-speech model, such as Kokoro 82M, whose weights you download and run on your own servers. There is no per-character fee, but you pay for the hardware.
