cancel_job
ChatGPTCancel a task that has not finished yet.
create_cue
ChatGPTAdd a cue to an existing paragraph and save a new script version. Read list_cues for paragraph indices.
create_profile
ChatGPTCreate a speaker profile for script context. Does not clone or generate a voice.
create_project
ChatGPTCreate an empty video project in the open PromptFy browser, optionally linking a saved script and profile. Does not generate or import media. Follow its job.
create_script
ChatGPTCreate a teleprompter script, optionally associated with an owned active profile. This stores a link; it does not rewrite supplied text.
delete_cue
ChatGPTRemove a cue and save a new version. Ask first, then pass confirm=true; script text remains unchanged.
delete_profile
ChatGPTSoft-delete a speaker profile. Ask first, then pass confirm=true. Scripts remain.
delete_project
ChatGPTSoft-delete the synced project and remove its local document in the running browser. Keeps local source media and scripts. Other devices may retain offline copies. Nothing is deleted unless confirm=true. Follow the job.
delete_script
ChatGPTDelete a script. Nothing is deleted unless confirm=true.
duplicate_project
ChatGPTDuplicate a project document with a new ID in the open browser. Media references are shared, not copied/uploaded. Close its editor first; follow the job.
edit_video
ChatGPTAutomatic edit on the user's computer (PromptFy web tab): transcribe and add captions, cut long pauses, J-cuts, punch-in zooms, smooth zooms, dissolves. Transcription uses the plan's monthly minutes. By default the result is saved as a copy of the project; in_place=true changes the original. j_cuts and dissolve can't be combined; neither can punch_in and smooth_zoom.
estimate_duration
ChatGPTEstimate how long the text takes to read aloud (default 150 words per minute).
export_video
ChatGPTExport a project to MP4 on the user's computer (PromptFy web tab). The file is downloaded there; nothing is uploaded. Returns name, length and size.
format_for_teleprompter
ChatGPTReflow text for reading on a teleprompter: one sentence per line, short lines, blank line between paragraphs. Does not save anything.
get_account
ChatGPTThe user's PromptFy plan.
get_audio_mixer
ChatGPTOpen or refresh the sound mixer for an identify_sounds job. Compatible hosts show percentage sliders; others receive the same state as text. This is read-only and can follow pending analysis/mix jobs without the devices:control scope. A mix_job_id must belong to this exact analysis.
get_job
ChatGPTStatus of a task running on the user's device.
get_profile
ChatGPTRead one speaker profile owned by the user.
get_project
ChatGPTInspect the synced project timeline: tracks, clip IDs, timings and associations. Does not read local files or prove that a device has the media.
get_project_transcript
ChatGPTWhat is said in a project's main video, in timeline order, plus the long pauses (in seconds). Empty until the project was transcribed (edit_video with captions).
get_script
ChatGPTGet one script with its full text.
identify_sounds
ChatGPTDetect sounds locally (up to 5 source minutes). Returns only detected class ids, names, confidence and time ranges. These are detections, not guaranteed separated stems. The caller chooses which sounds to remove or how much to retain (0–100%). Then call mix_audio with this job_id as analysis_id. device selects the source browser by ID or label from list_devices; required when multiple browsers are online.
isolate_voice
ChatGPTExtract speech locally (up to 10 source minutes). Downloads a separate WAV and saves a project copy for listening, preserving the original. No upload or voice credits. Returns file metadata and a project link. If available, use send_audio_to_chat(job_id=...) after completion to play a requested preview. Separation is an estimate; listen before using it. device selects the source browser by ID or label from list_devices; required when multiple browsers are online.
list_audio_clips
ChatGPTAudio-bearing clips in a project. Use their clip_id with isolate_voice or identify_sounds. Reads synced metadata only; source files stay local.
list_cues
ChatGPTList a script's web teleprompter cues. Paragraph indices are zero-based, split by blank lines.
list_devices
ChatGPTDevices where PromptFy can run tasks (the PromptFy web tab), and whether each one is open right now.
list_profiles
ChatGPTList the user's active speaker profiles, including writing tone and niche.
list_projects
ChatGPTThe user's video projects in the PromptFy editor, newest first: name, length, number of clips, and whether they already have a transcript or captions.
list_scripts
ChatGPTList the user's teleprompter scripts, newest first. Optional query filters by title or text. Returns id, title, word count and last update.
mix_audio
ChatGPTMix locally using a completed identify_sounds job. levels maps returned sound ids to percentages retained: 0 removes, 100 keeps. Omitted sounds and 'other' (the residual) default to 100. Levels are chosen by the caller. Downloads a WAV and saves a project copy; never changes the original. Estimates can overlap or contain leakage; inspect the result. The full audio stays in the source browser; send_audio_to_chat can share a requested preview. A changed source requires a fresh analysis.
open_in_teleprompter
ChatGPTOpen a script in the teleprompter on the user's PromptFy web tab, ready to read. device optionally names the device (from list_devices).
rename_project
ChatGPTRename a project in the open browser. Close its editor first. Follow the job.
reorder_project_clip
ChatGPTMove a clip within its magnetic timeline track to a zero-based index, using the editor's reorder rules (including linked audio). Read get_project first. Close the editor first; locked tracks and stale local documents are rejected.
send_audio_to_chat
ChatGPTPlay the requested audio in this chat. Returns up to 30 source seconds; longer files return a preview. The full file stays in the PromptFy project. After isolate_voice or mix_audio completes, pass its job_id to play the processed result; its source browser must remain open. Alternatively pass project_id and clip_id; omit clip_id only for a single audio-bearing clip. Source audio is sent without timeline effects. Audio passes through RAM, without server audio storage. Only share audio the user requests to hear here. This tool does not change the source file or use voice-generation credits.
set_project_links
ChatGPTAssociate a project's primary script/profile without changing media or text. Omitted links stay unchanged; use clear_script/clear_profile to detach them. Does not change script links already stored on individual recorded takes.
set_script_profile
ChatGPTAssociate a script with an owned active profile. Pass null to remove the association. Does not rewrite text.
update_cue
ChatGPTUpdate a cue and save a new version. Omitted fields, including appearance, are preserved.
update_profile
ChatGPTUpdate supplied profile fields only. Does not change its voice assignment.
update_script
ChatGPTChange a script's title and/or full text. content replaces the whole text.
wait_for_job
ChatGPTWait (up to 25 s) for a task on the user's device to finish, then return it.
cancel_job
Claudecreate_cue
Claudecreate_profile
Claudecreate_project
Claudecreate_script
Claudedelete_cue
Claudedelete_profile
Claudedelete_project
Claudedelete_script
Claudeduplicate_project
Claudeedit_video
Claudeestimate_duration
Claudeexport_video
Claudeformat_for_teleprompter
Claudeget_account
Claudeget_audio_mixer
Claudeget_job
Claudeget_profile
Claudeget_project
Claudeget_project_transcript
Claudeget_script
Claudeidentify_sounds
Claudeisolate_voice
Claudelist_audio_clips
Claudelist_cues
Claudelist_devices
Claudelist_profiles
Claudelist_projects
Claudelist_scripts
Claudemix_audio
Claudeopen_in_teleprompter
Clauderename_project
Claudereorder_project_clip
Claudesend_audio_to_chat
Claudeset_project_links
Claudeset_script_profile
Claudeupdate_cue
Claudeupdate_profile
Claudeupdate_script
Claudewait_for_job
Claude