Resources · Zapier integration

AssemblyAI Zapier integrationAutomate AssemblyAI with Zapier.

The AssemblyAI Zapier integration gives you 0 triggers and 9 actions. It sends recordings to AssemblyAI for speech-to-text, then pulls back the transcript, subtitles, paragraphs or an AI answer about it. Here is what each step does and costs.

See the 9 operations

Verified Trustpilot reviews · AI, automation & growth agency

Overview

What the AssemblyAI Zapier integration actually does

AssemblyAI is a speech-to-text service: you give it an audio or video file and it returns the words spoken, with extras such as speaker labels or a summary. Zapier carries those files in from your other apps and sends the text back out, without code. The result is a recording that becomes searchable, quotable text without anyone typing it.

Three results people build with it. Meeting or call notes: Transcribe with speaker labels and summarization switched on, then the text posted to your notes tool. Subtitles for every video: Transcribe, then Get Subtitles in SRT or VTT, ready for your video player. Answers about a recording: Prompt Transcript With LLM Gateway sends a prompt about the meeting to a language model and returns its answer, such as a short recap.

What it does not do: AssemblyAI cannot start a Zap. It has no trigger, so the recording always arrives from another tool, and nothing on this side tells Zapier when a transcript is ready. There is no search step either; to reach a transcript later, you keep its ID from the step that created it and pass it along.

The same app is covered on the AssemblyAI n8n integration and AssemblyAI Make integration pages. Counts here come from Zapier's app page, read on 1 October 2026 (AssemblyAiCLIAPI@3.0.0).

Vocabulary

Zapier vocabulary in one minute

Seven words cover everything below. Each one is explained once, then used as is.

Zap
The automation you build: one starting event, then one or more steps that run every time that event happens.
Trigger
The event in an app that starts the Zap. Once the Zap is on, Zapier waits for it. AssemblyAI offers none of its own.
Action
What the Zap does after it starts: the step that sends, fetches or deletes something for you, here a transcript.
Task
The unit Zapier bills: one action that completed successfully. Your plan's allowance is a count of these.
Polling vs instant
A polling trigger is checked on a timer set by your plan. An instant one is reported by the app the moment it happens.
Filter
A step that lets the Zap continue only when the incoming data meets your condition, and stops it otherwise.
Multi-step Zap
Any Zap that goes beyond one trigger and one action. A second action or a filter makes one, and it needs a paid plan.
Cost

What it really costs to run

AssemblyAI is not a premium app on Zapier, so it works on the free plan. The real limit is how many AssemblyAI steps you chain.

  • Free plan100 tasks a month, unlimited Zaps
  • Free plan Zap size2 steps: one trigger, one action
  • Billedeach successful action
  • Not billedthe trigger, checks for new data, failed actions

Each action that succeeds counts as one task. Transcribe followed by Get Subtitles uses two tasks per recording; add a prompt and it is three. A failed action is not counted, and the trigger never costs anything. The free plan stops at two steps: one trigger and one action. A Zap that chains two AssemblyAI actions, or adds a filter, becomes multi-step and needs a paid plan. Plan for that before you build.

Connect

Connect AssemblyAI to Zapier in 3 steps

  1. 01

    Open your connections

    In Zapier, go to the Apps page of your account and click + Add connection. A connection is your AssemblyAI account linked once, then reused by every step that needs it.

  2. 02

    Pick AssemblyAI

    Type AssemblyAI in the search box of the dialog, select it, then click Add connection. A new tab opens on the AssemblyAI side, where you finish linking the account.

  3. 03

    Sign in and allow access

    Log in with the AssemblyAI account that should own the transcripts, and grant Zapier access. Back in Zapier, the account appears in the list and every AssemblyAI step can pick it.

First Zap

Your first Zap: subtitles for every new recording

GoalEach new recording is transcribed by AssemblyAI and comes back as a subtitle file, ready for your video player.

  1. 01

    Create the Zap and pick the starting event

    Start a new Zap and choose the new-recording event of the tool where your files land. Connect it and test: Zapier brings back the three most recent recordings so you work on a real one.

  2. 02

    Add Transcribe

    Add an AssemblyAI step with Transcribe. Map the file link into Audio URL, switch on the options you need, and turn on Wait for Completion so the step returns a finished transcript.

  3. 03

    Add Get Subtitles

    Add a second AssemblyAI step, Get Subtitles. Map the transcript ID from the step above and pick SRT or VTT. This second action makes the Zap multi-step, so it needs a paid plan.

  4. 04

    Test both steps

    Run each step once on the sample recording. Read what each one returned: the transcript for names and jargon, then the subtitle output. A wrong file link shows up at Transcribe, not later.

  5. 05

    Publish

    Click Publish to switch the Zap on. Each recording then costs two tasks, one per AssemblyAI action. The trigger costs nothing.

Actions

The 9 AssemblyAI actions

AssemblyAI gives you 9 actions. For each one: what it does for you, when to reach for it, and what to watch out for.

1

Delete Transcript

Action

In Zapier“Remove the data from the transcript and mark it as deleted.”

Erases the content of a transcript on AssemblyAI and marks it as deleted. Only the transcript ID is needed.

When to use it
recordings with personal data that should not stay stored once the text has been copied elsewhere.
Watch out
run it last, after every step that still reads the transcript.
2

Get Paragraphs

Action

In Zapier“Get the transcript split by paragraphs. The API will attempt to semantically segment your transcript into paragraphs to create more reader-friendly transcripts.”

Returns the same transcript cut into paragraphs that follow the meaning of the speech, which reads far better than one block of text.

When to use it
turning an interview or a podcast episode into an article draft or a blog post.
Watch out
the transcript ID is required, so this step comes after one that provides it, usually Transcribe.
3

Get Sentences

Action

In Zapier“Get the transcript split by sentences. The API will attempt to semantically segment the transcript into sentences to create more reader-friendly transcripts.”

Splits a finished transcript sentence by sentence, an easier unit to quote than whole paragraphs.

When to use it
pulling quotable lines from a customer call for a testimonial review or a sales training sheet.
Watch out
it needs the transcript ID from an earlier step.
4

Get Subtitles

Action

In Zapier“Export your transcript in SRT or VTT format to use with a video player for subtitles and closed captions.”

Exports a transcript as a subtitle file in SRT or VTT, the two formats video players read. Both the transcript ID and the format are required.

When to use it
captions for every webinar replay or course video you publish.
Watch out
check which of the two formats your player or platform expects before publishing.
5

Get Transcript

Action

In Zapier“Get the transcript resource. The transcript is ready when the “status” is “completed”.”

Fetches a transcript you created earlier, by its ID. It is ready once its status reads completed.

When to use it
a later Zap that picks up a transcript started without waiting.
Watch out
called too soon, it returns an unfinished job.
6

Prompt Transcript With LLM Gateway

Action

In Zapier“Generates a response from a model given a prompt or a series of messages.”

Sends a prompt, or a series of messages, to a language model through AssemblyAI and returns the answer. The Model field is required.

When to use it
asking for a recap, the decisions taken or the next steps of a meeting, then posting them to your team chat.
Watch out
the Model field is required, so pick the model before testing the step.
7

Search Words

Action

In Zapier“Search through the transcript for keywords. You can search for individual words, numbers, or phrases containing up to five words or numbers.”

Looks inside one transcript for words, numbers or short phrases of up to five words, so a Zap can act on calls that mention them.

When to use it
flagging every call where a competitor or a cancellation is mentioned, then sending it to a manager.
Watch out
despite its name, it is an action and costs a task.
8

Process Speech Understanding

Action

In Zapier“Perform various speech understanding tasks like translation, speaker identification, and custom formatting.”

Runs extra processing on a finished transcript, such as translation, naming the speakers or custom formatting.

When to use it
a sales call where the speakers should carry the rep's and the client's names instead of generic labels.
Watch out
the request field expects JSON, the least beginner-friendly field of the integration.
9

Transcribe

Action

In Zapier“Submit an audio file for transcription”

Sends an audio or video file to AssemblyAI and starts the speech-to-text job. Only the file link is required; options add speaker labels, language detection, summaries, sentiment, chapters or redaction of personal data.

When to use it
the first AssemblyAI step of almost every Zap.
Watch out
with Wait for Completion off, you only get an ID back, not the text.
When it breaks

When it breaks

Zapier lists no help article for AssemblyAI, so the usual failures come from how Zaps work in general, plus the timing of transcription.

Most problems show up as a step that returns no text, or a Zap that runs far more than planned. For step-by-step fixes, the Zapier troubleshooting guide goes further.

  • Transcript not finished yet

    Transcribe waits up to 15 minutes when Wait for Completion is on. With it off, the next steps get an ID and no text, so they fail or return nothing.

  • No backfill of old recordings

    A Zap only reacts to what happens after you publish it. Recordings saved before that will not be transcribed; that backlog is a separate job.

  • The loop that eats your tasks

    If the Zap saves its output into the very folder or list its trigger watches, each run sets off the next one and the task count drains fast.

  • Bulk uploads held back

    When the trigger app suddenly returns a huge batch, after an import or a migration, Zapier's flood protection can hold the runs back.
Need help

Need help automating AssemblyAI with Zapier?

A person reads every message.

FAQ

AssemblyAI Zapier integration: FAQ

01Is the AssemblyAI Zapier integration free?
On the Zapier side, yes for two-step Zaps. AssemblyAI is not a premium app, so the free plan accepts it, with 100 tasks a month and unlimited Zaps. The catch is the shape: one trigger from another app plus one AssemblyAI action. Many Zaps chain Transcribe with Get Subtitles, Get Paragraphs or a prompt, and that second action makes them multi-step, which needs a paid plan. The trigger from the other app never costs a task, and a failed action is not counted.
02How many tasks does transcribing a file use?
One task per successful action. Transcribe alone uses one task per recording. Transcribe plus Get Subtitles uses two, and adding a prompt brings it to three, since every action that succeeds is counted on its own. The trigger never costs a task, Zapier checking for new recordings costs nothing, and a failed action is not billed. Keep the chain short: every extra AssemblyAI step you add is one more task per recording.
03Does it work in real time?
Not in the live-captioning sense. AssemblyAI has no trigger, so the Zap starts on the other app's event: instant if that trigger is instant, otherwise on a timer of 15 minutes on Free, 2 on Professional and 1 on Team and Enterprise. Then the transcription itself takes time. With Wait for Completion on, Transcribe waits up to 15 minutes for the result; with it off, it returns at once with only the transcript ID, and the text has to be fetched later.
04What do you need to connect AssemblyAI?
An AssemblyAI account, and recordings that sit at a link AssemblyAI can open. The connection itself takes three moves: open Apps in Zapier, click + Add connection, pick AssemblyAI, then sign in and grant access in the tab that opens. The Audio URL field of Transcribe is required, so each recording has to come with a link the Zap can map. If the recordings belong to a colleague, you also need access to them where they are stored.
05Why does Get Transcript come back without text?
Because the transcript is not finished yet. AssemblyAI marks a transcript as ready only when its status reads completed, and Get Transcript returns whatever state the job is in. This happens when Transcribe ran with Wait for Completion switched off, so it handed back an ID straight away. The simplest fix is to switch that option on, so Transcribe waits up to 15 minutes. Otherwise, fetch the transcript in a later Zap, once enough time has passed.
06What if the operation you need is not in the list?
Then the integration does not offer it: the visible list is 9 actions, with no trigger and no search. Process Speech Understanding covers several extra tasks on its own, such as translation and speaker naming, so check it first. Get Transcript, Get Paragraphs and Get Sentences also return the same transcript in different cuts, which may already cover the need. Beyond that, the same app on Make or n8n exposes its own set of modules, which may fit a specific need better.
07Zapier, Make or n8n for AssemblyAI?
It depends on your criteria, not on a winner. On Zapier, the deciding facts are on this page: 9 actions, no trigger and no search on the AssemblyAI side, a Transcribe step that waits up to 15 minutes, and billing per successful action, so Transcribe plus Get Subtitles costs two tasks. Check that the action you need is in that list, then open the same app on Make or n8n and compare what each exposes and how it counts a run. The tool your team already uses weighs in too.
Same platform

Other docs you might need on this stack.

Browse all integrations