Rates and claims last checked August 2026
Paxa Labs and Narakeet for Thai speech
A production tool and a speech API compared on what each is shaped to do, how they bill, and which one belongs inside your software.
Narakeet is built around producing voiceovers and narrated video from a script, with an API beside that workflow. We do not make video and have no plans to. What we sell is the speech layer of somebody else's product: a call your software makes while a user waits, or while an LLM is still writing the sentence being spoken. Those are different jobs, and a team that needs the first one should not buy the second.
What differs
| Dimension | Paxa Labs | Narakeet |
|---|---|---|
| What the product is shaped to do | Serve speech inside your application, request by request, including while text is still being generated. | Produce voiceovers and narrated video from a script, with an API beside that workflow. |
| Real-time delivery | Chunked HTTP streaming and a live websocket. Audio starts before the full text exists. | Their documentation is where to check what real-time delivery they offer. |
| Word timestamps | Word, sentence, or utterance spans on the same call, at a stated multiplier on the per-character rate. | Their documentation is where to check what timing data they return. |
| Thai voices | 26 in the roster, including four regional registers, each with a published id and a fixed sample line. | Their voice list, including its Thai entries, is on their site. |
| What a failed request costs | Nothing. Credits are charged before inference and refunded automatically when generation fails, including mid-stream. | Their billing terms are on their pricing page. |
- Paxa Flash, per 1M characters
- $15
- Narakeet
- $70-280
- Their rate, as a multiple
- 4.7-19×
Both figures are published list rates normalized to one million characters, read from each vendor's own page and re-checked on a fixed schedule.
What it costs at your volume
| Speech, per month | Paxa Labs | Narakeet | Difference |
|---|---|---|---|
| 100,000About 1.8 hours of speech | $1.50 | $7-28 | $5.50 less |
| 1,000,000About 18 hours | $15 | $70-280 | $55 less |
| 5,000,000About 92 hours | $75 | $350-1,400 | $275 less |
| 20,000,000About 370 hours | $300 | $1,400-5,600 | $1,100 less |
Every figure is the two published rates multiplied by the volume beside it. Where the other vendor publishes a range, the low end is compared and the range is shown.
When Narakeet is the better choice
If the output is a finished video or a narrated presentation, a tool that produces one is worth more than an API that returns an audio file. Narakeet has built that workflow and we have not. Reassembling it out of a speech endpoint and a video library is a project, and buying it is a purchase. For a content team that difference decides the question on its own. Choose us when the speech goes into software you are already building.
Questions
- Do you produce video?
- No. We return audio, timing spans, or a live stream of audio. Everything that happens to it afterwards is yours.
- Can we generate long narration?
- Each request takes up to 5000 characters. Long scripts are sent in consecutive requests, at the same per-character rate, with no per-request minimum beyond the small floor.
- Do you publish quality comparisons?
- No. We publish no quality or accuracy figures of our own, for our model or anyone else's. Generate the same script on both and listen.
- How current are the prices here?
- They come from each vendor's published rates, recorded in one place and re-checked on a fixed schedule; the month of the last check is at the top of the page.
Start with free credits
Sign in with Google, GitHub, or Hugging Face and spend 100 one-time free credits on your own text.