First-time listener
You want to hear whether an AI voice suits a short greeting.
Test a brief sentence and check whether playback, saving, or repeated attempts have different conditions.
Access guide
The speech-to-text tool queues text-to-speech jobs, clones voices from a short reference clip, and transcribes audio. New accounts receive 1,000 credits usable without a subscription; it does not offer real-time voice agents or talking avatars.
If you are asking is fish audio free, verify the current terms rather than relying on an old screenshot or review. Start with a short voice script, then check what the service actually lets you generate, replay, and use.
Illustrative example: Write a short voice script
Illustrative voice script only. Nothing is submitted here; enter your text or upload your reference clip in the speech-to-text tool after continuing.
Bring a short script and decide what you need to do with the resulting audio. A test for personal listening may not answer the same access question as a published project.
A usable test is not necessarily permission for every later use. Match your access check to the job you have in mind.
You want to hear whether an AI voice suits a short greeting.
Test a brief sentence and check whether playback, saving, or repeated attempts have different conditions.
You need narration for a clip you may publish.
Check the current usage terms before treating a successful voice test as approval to publish.
You expect to revise scripts and generate several versions.
Check whether the access available to you covers the number of attempts and output use your project requires.
Use the same small test from start to finish so you can separate voice quality from access restrictions.
On the service you intend to use, look for the present access conditions and any stated restrictions on generation, downloads, or use of output. Do this before entering a long script; terms can change.
Enter two sentences you wrote or have permission to use. Generate a voice sample if the available access permits it, then listen for pronunciation, pacing, and whether another attempt is needed.
See whether you can perform the action your project needs, such as revising the script or using the audio in a published clip. If a condition is unclear, do not assume that hearing a sample grants broader rights.
Compare these two questions separately: can you test a voice, and can you use the result for your intended project?
These illustrations show a decision process, not confirmed access tiers or a guarantee that output can be downloaded or published.
Test a short scriptCheck intended useAt the speech-to-text tool, signup gives 1,000 credits usable without a subscription. Credits can run out, and output-use terms should still be checked.
A past mention of free access does not establish what is available on the day you visit.
What to do instead
Read the terms shown by the service you actually open.
Being able to hear generated speech does not, by itself, settle whether you may use it in a public or commercial project.
What to do instead
Check the applicable output-use terms before publishing.
Longer scripts, repeated revisions, or saving output may be subject to conditions that a brief trial does not reveal.
What to do instead
Test the specific action your workflow depends on before preparing all your copy.
Open the speech-to-text tool to try a short passage with your 1,000 signup credits. You need no subscription to use that balance; check the current output terms before using any recording publicly.
The partner link opens the speech-to-text tool, not Fish Audio. New speech-to-text tool accounts get 1,000 credits usable without a subscription; use can exhaust them. Check the current output-use terms before publishing.
A short script is a practical way to assess a voice if testing is available under the current terms. Read those terms first, then use a passage you wrote or have permission to use.
No conclusion about publication rights follows from playback alone. Review the applicable output-use conditions for your intended destination before posting or distributing audio.
List the actions you need, such as revising a script, generating multiple takes, saving audio, and publishing the result. Check each action against the current terms rather than judging access from one successful test.