Compare · Saidwire vs Speechmatics

Saidwire vs Speechmatics for voice agents

On Coval's public benchmark, Wire-1 by Saidwire returns the final transcript 35 ms on average after the end of speech (Coval, last measured Nov 10, 2026), and Speechmatics's quickest listed model, Linden 1, takes 235 ms (Coval, last measured Nov 10, 2026). On error rate the two are close: Wire-1 measures 3.7% and Linden 1 measures 3.8% (Coval, last measured Nov 10, 2026). Choose Speechmatics if you want turn detection from the speech-to-text vendor, or you need many languages; choose Saidwire if the wait before your agent can answer matters most and your callers speak English.

Speed and accuracy

Every Speechmatics speech-to-text model Coval lists, next to Wire-1. Lower is better in both columns.

Listed price

Where Speechmatics is the better choice

• Built-in turn detection: its pricing page describes Linden 1 as a voice agent model with built-in turn detection. With Saidwire you bring your own turn detector or use a pause setting. (Speechmatics pricing page, read Nov 10, 2026)
• Languages: the same pricing page lists a multilingual model covering 56 languages. Wire-1 is English only. (Speechmatics pricing page, read Nov 10, 2026)
• Linden 1 is listed at $0.15 per hour, below Saidwire's $0.25 list price. (Speechmatics pricing page, read Nov 10, 2026)

Where Saidwire is the better choice

• Time to final transcript: 35 ms on average and 50 ms at the 95th percentile, 1st of 29 (Coval, last measured Nov 10, 2026). Linden 1 from Speechmatics measures 235 ms on average (Coval, last measured Nov 10, 2026).
• A long free trial: 120 audio hours with no card and no deadline, on up to 4 streams at once.
• Audio is processed in memory and never stored, and transcripts are not logged (security and data).

Switching

If your agent runs on Pipecat, LiveKit Agents or Vapi, switching is a one-line change: see the Pipecat, LiveKit and Vapi guides. With your own client, connect to wss://api.saidwire.com/v1/listen.

Test both on your own audio

Stream the same call through each API and time the gap between the end of speech and the final result. The Saidwire side: Get a free Saidwire key at saidwire.com/dev.

# pip install websockets   (Python 3.10+)
import asyncio, json, os
from websockets.asyncio.client import connect

URL = ("wss://api.saidwire.com/v1/listen?encoding=linear16&sample_rate=16000"
       "&interim_results=true&punctuate=true&endpointing=false")

async def run(frames):                     # frames: 20-100 ms chunks of 16-bit PCM
    headers = {"Authorization": f"Token {os.environ['SAIDWIRE_API_KEY']}"}
    async with connect(URL, additional_headers=headers) as ws:
        async def read():
            async for raw in ws:
                m = json.loads(raw)
                if m.get("type") == "Results":
                    text = m["channel"]["alternatives"][0]["transcript"]
                    print("final  " if m["is_final"] else "interim", text)
        reader = asyncio.create_task(read())
        async for chunk in frames:
            await ws.send(chunk)
        # your turn detector says the user stopped: ask for the final now
        await ws.send(json.dumps({"type": "Finalize"}))
        await ws.send(json.dumps({"type": "CloseStream"}))
        await reader

Test Speechmatics with your own Speechmatics account and its SDK, under its terms. We do not run or publish tests of other vendors.

How to read these numbers

• Who measured. Speed and error-rate figures are Coval's, from its public benchmark pages. Coval runs every model daily on the same audio and starts every timer from one shared end-of-speech point. Saidwire ran none of these tests and does not test other vendors.
• Time to final segment is the wait between the end of speech and the final transcript: the wait a voice agent has before it can answer. It leaves out your network round trip and each vendor's own turn-detection delay.
• Word error rate is the share of words a transcript gets wrong on Coval's test audio. Your audio will differ.
• Prices are list prices read from each vendor's public pricing page on Nov 10, 2026. Plans, discounts and add-ons change them; the vendor's page is the authority.
• What Coval leaves out. Coval does not score every model, language, add-on or setting a vendor offers, and it says so on each provider page. Features such as speaker labels, translation and language coverage are not part of these two numbers.
• Figures roll daily. Each row links to the page it came from, with that page's date.

Questions

Questions or stuck? Email support@saidwire.com. Saidwire LLC, 215 W Pershing Rd Ste 403 PMB 1093, Kansas City MO 64108.

Speechmatics and the other names on this page belong to their respective owners. They are named only to describe compatibility or to compare public figures; no partnership or endorsement is implied. Coval benchmark figures are from Coval's public pages; Coval does not endorse Saidwire.