← FreeTTSCONTENT HISTORY

Update to FreeTTS

Snapshot Oct 31, 2026 · 00:34 UTC · version 1.1.0

Collection source: downloaded plugin package.

WHAT CHANGED

Instructions updated for freetts-audio

Skill instructions — Now says: In ChatGPT the audio also plays in a FreeTTS player under your reply, with buttons to hear the same text in other voices. Keep your reply short: one line...

Observed in instructions or declared skills. Runtime behavior has not been tested.

Save FreeTTS and check future changes in Following.

Up to 3 plugins free · no card required.

All fields and original values.

Skill instructions

Before

Oct 2, 2026 · 18:02 UTC

No removed lines (ordering may have changed)

After

Oct 31, 2026 · 00:34 UTC

In ChatGPT the audio also plays in a FreeTTS player under your reply, with buttons to hear the same text in other voices. Keep your reply short: one line with the voice and length, plus the link.

Compare saved observations

Download comparison JSON

SKILL.md text changes

− removed · + added

--- before
+++ after
@@ -7,6 +7,8 @@
 
 FreeTTS reads text aloud in 149 languages with 2,253 voices and returns a link to the audio file. Always give the person the link, the voice used and the length.
 
+In ChatGPT the audio also plays in a FreeTTS player under your reply, with buttons to hear the same text in other voices. Keep your reply short: one line with the voice and length, plus the link.
+
 ## Pick the right tool
 
 - Plain text read aloud: `text_to_speech`. Pass `language` when no voice is named; FreeTTS picks a natural voice for it.
Full technical diff · 1 changed fields

changed /skill_md_contents

BEFORE
"---\nname: freetts-audio\ndescription: Turn text into speech with FreeTTS and hand back an MP3 link. Use when someone wants text read aloud, narrated, voiced, saved as audio, turned into a dialogue or podcast, timed as a voiceover or prompter track, or transcribed from an audio link.\n---\n\n# Making audio with FreeTTS\n\nFreeTTS reads text aloud in 149 languages with 2,253 voices and returns a link to the audio file. Always give the person the link, the voice used and the length.\n\n## Pick the right tool\n\n- Plain text read aloud: `text_to_speech`. Pass `language` when no voice is named; FreeTTS picks a natural voice for it.\n- The person has not chosen a voice: `suggest_voice` with the language and what the audio is for, then use its first pick.\n- \"What voices do you have for ...\": `list_voices` with the language, or read the resource `freetts://voices/{language}`.\n- Two or more speakers (a dialogue, interview, lesson or podcast): `dialogue_to_speech`, one line per speaker as `Name: text`. Pass `voices` when the person names them.\n- Exact timing (a presenter's ear prompter, a voiceover timed to video, a workout or recipe with pauses, rehearsal tracks): `script_to_tracks`. Read the resource `freetts://script-mode/syntax` first; `(pause 2)` is exactly two seconds, `(beep)` and `(ding)` are sounds, `TRACK: name` starts a new file, `CUE:` lines use a second voice.\n- A recording to text: `transcribe_audio` with a public https link to the audio file.\n- Questions about limits or what an account can do: `check_usage`, or the resource `freetts://plans`.\n\n## Good results\n\n- Keep the person's words exactly as written unless they ask for edits. Split long text at paragraph breaks if it is longer than one request allows, and say so.\n- For names or unusual words, spell them the way they should sound.\n- Match the voice to the language of the text; do not read Spanish text with an English voice.\n- For children, lessons and audiobooks, a calm, clear voice at normal speed works best; `speed` from -15 to -5 helps learners.\n\n## Accounts\n\n- Some tools need a connected FreeTTS account. When a tool says so, tell the person plainly in one sentence and share the link it gives, and leave it at that.\n"
AFTER
"---\nname: freetts-audio\ndescription: Turn text into speech with FreeTTS and hand back an MP3 link. Use when someone wants text read aloud, narrated, voiced, saved as audio, turned into a dialogue or podcast, timed as a voiceover or prompter track, or transcribed from an audio link.\n---\n\n# Making audio with FreeTTS\n\nFreeTTS reads text aloud in 149 languages with 2,253 voices and returns a link to the audio file. Always give the person the link, the voice used and the length.\n\nIn ChatGPT the audio also plays in a FreeTTS player under your reply, with buttons to hear the same text in other voices. Keep your reply short: one line with the voice and length, plus the link.\n\n## Pick the right tool\n\n- Plain text read aloud: `text_to_speech`. Pass `language` when no voice is named; FreeTTS picks a natural voice for it.\n- The person has not chosen a voice: `suggest_voice` with the language and what the audio is for, then use its first pick.\n- \"What voices do you have for ...\": `list_voices` with the language, or read the resource `freetts://voices/{language}`.\n- Two or more speakers (a dialogue, interview, lesson or podcast): `dialogue_to_speech`, one line per speaker as `Name: text`. Pass `voices` when the person names them.\n- Exact timing (a presenter's ear prompter, a voiceover timed to video, a workout or recipe with pauses, rehearsal tracks): `script_to_tracks`. Read the resource `freetts://script-mode/syntax` first; `(pause 2)` is exactly two seconds, `(beep)` and `(ding)` are sounds, `TRACK: name` starts a new file, `CUE:` lines use a second voice.\n- A recording to text: `transcribe_audio` with a public https link to the audio file.\n- Questions about limits or what an account can do: `check_usage`, or the resource `freetts://plans`.\n\n## Good results\n\n- Keep the person's words exactly as written unless they ask for edits. Split long text at paragraph breaks if it is longer than one request allows, and say so.\n- For names or unusual words, spell them the way they should sound.\n- Match the voice to the language of the text; do not read Spanish text with an English voice.\n- For children, lessons and audiobooks, a calm, clear voice at normal speed works best; `speed` from -15 to -5 helps learners.\n\n## Accounts\n\n- Some tools need a connected FreeTTS account. When a tool says so, tell the person plainly in one sentence and share the link it gives, and leave it at that.\n"
Full snapshot data
{
  "description": "Turn text into speech with FreeTTS and hand back an MP3 link. Use when someone wants text read aloud, narrated, voiced, saved as audio, turned into a dialogue or podcast, timed as a voiceover or prompter track, or transcribed from an audio link.",
  "included_files": [],
  "name": "freetts-audio",
  "skill_md_contents": "---\nname: freetts-audio\ndescription: Turn text into speech with FreeTTS and hand back an MP3 link. Use when someone wants text read aloud, narrated, voiced, saved as audio, turned into a dialogue or podcast, timed as a voiceover or prompter track, or transcribed from an audio link.\n---\n\n# Making audio with FreeTTS\n\nFreeTTS reads text aloud in 149 languages with 2,253 voices and returns a link to the audio file. Always give the person the link, the voice used and the length.\n\nIn ChatGPT the audio also plays in a FreeTTS player under your reply, with buttons to hear the same text in other voices. Keep your reply short: one line with the voice and length, plus the link.\n\n## Pick the right tool\n\n- Plain text read aloud: `text_to_speech`. Pass `language` when no voice is named; FreeTTS picks a natural voice for it.\n- The person has not chosen a voice: `suggest_voice` with the language and what the audio is for, then use its first pick.\n- \"What voices do you have for ...\": `list_voices` with the language, or read the resource `freetts://voices/{language}`.\n- Two or more speakers (a dialogue, interview, lesson or podcast): `dialogue_to_speech`, one line per speaker as `Name: text`. Pass `voices` when the person names them.\n- Exact timing (a presenter's ear prompter, a voiceover timed to video, a workout or recipe with pauses, rehearsal tracks): `script_to_tracks`. Read the resource `freetts://script-mode/syntax` first; `(pause 2)` is exactly two seconds, `(beep)` and `(ding)` are sounds, `TRACK: name` starts a new file, `CUE:` lines use a second voice.\n- A recording to text: `transcribe_audio` with a public https link to the audio file.\n- Questions about limits or what an account can do: `check_usage`, or the resource `freetts://plans`.\n\n## Good results\n\n- Keep the person's words exactly as written unless they ask for edits. Split long text at paragraph breaks if it is longer than one request allows, and say so.\n- For names or unusual words, spell them the way they should sound.\n- Match the voice to the language of the text; do not read Spanish text with an English voice.\n- For children, lessons and audiobooks, a calm, clear voice at normal speed works best; `speed` from -15 to -5 helps learners.\n\n## Accounts\n\n- Some tools need a connected FreeTTS account. When a tool says so, tell the person plainly in one sentence and share the link it gives, and leave it at that.\n"
}

SHA-256 of public snapshot: 38bcf35ff807b89697bb830f126d1b1626b54310b128969f42ed01bc937c3adf