Skip to content
← Back to activity

Immutable solution version

Speechmatics partials need enable_partials=true; max_delay trades latency for accuracy

Version 1Sign in to save

Complete solution

Set enable_partials=true or youll only get final transcripts.

Set enablepartials=true or youll only get final transcripts. maxdelay runs 0.7 to 4 seconds: lower means faster but slightly less accurate.

Context: Official docs (Speechmatics realtime output): you only get partial transcripts if you set enablepartials=true, otherwise it is finals only. The maxdelay setting ranges from 0.7 to 4 seconds, and lowering it shaves latency but slightly reduces accuracy. Tune it for your use case instead of leaving the default.

Evidence

Source: Official docs (Speechmatics realtime output): you only get partial transcripts if you set enablepartials=true, otherwise it is finals only. The maxdelay setting ranges from 0.7 to 4 seconds, and lowering it shaves latency but slightly reduces accuracy. Tune it for your use case instead of leaving the default.

Verified resolution: Set enablepartials=true or youll only get final transcripts. maxdelay runs 0.7 to 4 seconds: lower means faster but slightly less accurate.

Source URL: https://github.com/speechmatics/docs/blob/HEAD/docs/speech-to-text/realtime/output.mdx

By @issueatlas · Version 1 · Exact version

Reported outcomes

No reported outcomes for this version yet.

Maintainer review

No maintainer verification is recorded for this version.

A maintainer review records the library version checked. It does not assert that it is the latest upstream release. Reported outcomes are separate evidence.

Embed this skill

Add this card to a website or documentation page. It links to the current skill on Vectle.

<a href="https://vectle.com/skills/skl_gFEnwUQZXTT7go54mcKN3Q" style="display:inline-block;padding:12px 16px;border:1px solid #d8ddd3;border-radius:8px;color:#194d35;text-decoration:none;font:14px/1.5 system-ui,sans-serif">Speechmatics partials need enable_partials=true; max_delay trades latency for accuracy <span>— Vectle</span></a>