HALOWERK Medien

Takes an audio file (MP3, M4A, Ogg, FLAC; up to 24 MB). Without set it returns all metadata tags, embedded chapters and whether a cover image is present. With set it writes the given fields — title, artist, album, album_artist, genre, date, track, comment, composer, publisher, copyright, language, description — as ID3v2.3 for MP3 or native tags for other containers, removes fields set to an empty

Takes a video and an audio file (for example a dubbed track from /v1/dub, a voice-over or music) and returns the video with the new sound: mode replace drops the original audio, mode mix keeps it at original_volume underneath. The new audio is trimmed to the video length or padded with silence, optionally delayed by offset_seconds. The picture is copied, not re-encoded. Output MP4 with AAC inline

Takes a video of up to 15 minutes, measures the integrated loudness, true peak and loudness range of its soundtrack, and normalises it to the chosen standard with two-pass linear loudnorm, so the dynamics are kept and the target is met closely. The picture is copied without re-encoding. Loudness is measured again on the result and both measurements are returned. Output MP4 with AAC audio inline as

Takes an episode file (up to 24 MB) and checks it against widely used podcast delivery recommendations: MP3 or AAC container, bitrate of at least 64 kbit/s (mono) or 96 kbit/s (stereo), 44.1 or 48 kHz sample rate, integrated loudness of -16 LUFS ±1 for stereo (-19 for mono), true peak at most -1 dBTP, a title tag present, chapters present (recommendation only), and no more than 3 seconds of silenc

Takes up to 100 words, names or short phrases in one of seven languages and returns one MP3 per item spoken by a natural synthetic voice, optionally a second version at 75 % speed (pitch kept) for learners, and a combined file with all items in order separated by pauses. File names follow the list position. Ready to attach to flashcards (Anki, Quizlet) or a glossary. Speech by Deepgram with traini

Takes up to 30 prompts, each with an id and its text (total up to 10,000 characters), and returns one audio file per prompt in telephone format: 8 kHz mono WAV in 16-bit PCM, G.711 A-law (Europe) or µ-law (North America), or MP3 for preview. Files are named after the ids so they can be dropped into a phone system. Short silence is added at start and end, as phone systems expect; the voice is selec

Takes a text of up to 10,000 characters with simple markers: [pause N] inserts N seconds of silence (up to 30), and a blank line inserts the paragraph pause you set. Every text part is spoken by a natural synthetic voice in one of seven languages, and the parts and pauses are joined exactly, so a breathing exercise or a dictation gets the gaps it needs. Returned are the audio file (MP3, Opus or WA

Takes a script of up to 200 lines (together up to 12,000 characters), each with a speaker role and text, and returns one audio file: every role gets its own voice — assigned automatically from the voices of the chosen language, or set per role — and lines are separated by a pause you choose. Returned with the file are the start and end time of every line and an SRT, so the dialogue can be shown al

Takes a recording of up to 120 minutes and returns every instance of profanity, slurs, personal insults and explicit sexual or violent language, each with category, severity (mild, moderate, strong), the quote and timestamp, plus a summary for ad-suitability decisions. Context matters and is considered: quoting or discussing a word is marked as such.

Takes a recording and a topic or question (for example "pricing", "the new law", "why they founded the company") and returns up to the requested number of verbatim, self-contained statements about it that work as quotes in an article, post or report, each with timestamp, speaker label if available, and a short note on what it says.

Takes a recording of up to 120 minutes and returns the specialist terms, abbreviations and jargon it uses, each with a short explanation based on how it is used in the recording (not a general dictionary definition), its category, the quote where it first appears and the timestamp. Every finding carries a verbatim quote from the transcript; quotes are searched in the transcript and findings whose

Takes a recording of up to 60 minutes and returns a report of the personal data it contains: names, postal addresses, phone numbers, e-mail addresses, dates of birth, bank and card numbers, ID numbers, and special categories such as health information, each with category, a GDPR-oriented sensitivity rating (normal or special category), the quote and the timestamp. Nothing is redacted — use /v1/aud

Takes a recording of up to 120 minutes and a list of up to 30 brand, product or competitor names, and returns every mention of them — also in obvious variants and abbreviations — with the tone of the mention, a one-line context, whether it is a paid or sponsored mention, the quote and the timestamp, plus counts per name. Names not in the list are ignored.

Takes a recording of up to 60 minutes and derives FAQ entries from it: the questions that were actually asked (or clearly implied), rewritten as clean FAQ questions, with the answer as it was given in the recording, a topic tag, and the quote and timestamp where the answer is given. Questions that were asked but not answered are listed as open so the knowledge base can be completed.

Takes a recording of up to 60 minutes in the language being learned and a learner level, and returns words and phrases worth learning — chosen for the level, skipping the most basic ones — each with the base form, translation into the learner's language, word type, an estimated CEFR level, and the exact example sentence from the recording with its timestamp so it can be listened to in context.

Takes a recording of up to 30 minutes and creates listening-comprehension exercises like those in language exams: multiple-choice questions with four options and true/false statements, at the chosen CEFR level, in the learner's or the recording's language. Every question comes with the correct answer and the exact passage (quote and timestamp) where the answer is heard, so learners can replay it.

Takes a recording of up to 120 minutes and returns the statements in it that assert something checkable — statistics, dates, causal claims, comparisons, attributions — rewritten as a neutral claim, typed, and with a note on what source or data would settle it. Opinions, predictions and jokes are left out. It does not judge whether a claim is true; it prepares fact-checking.

Takes a call or meeting of up to 60 minutes, separates the speakers and returns every commitment made: what was promised, by which speaker, the deadline or condition as spoken, its kind (delivery, price, refund, callback, guarantee, other), the quote and timestamp. Useful for sales and support follow-up, dispute evidence and compliance reviews.

Takes a sales call or demo of up to 60 minutes, separates the speakers and returns every objection raised: its category (price, timing, competitor, authority, need, trust, other), the speaker, the quote, a short summary of how it was answered, and whether it was resolved, partly resolved or left open. For sales coaching and win-loss analysis.

Takes a voice note or dictation of up to 10 minutes ("write to Mr Weber that the offer is late and we will send it Friday...") and returns an e-mail draft: subject, salutation, body and closing, in formal or informal register and the chosen tone, with names, dates and figures exactly as spoken. Points the speaker left open are listed separately instead of being invented. Nothing is sent; the draft

Takes a recording of up to 120 minutes and the platforms you want, and writes posts for each: a LinkedIn post, an X post or short thread, an Instagram caption, and so on, each within the platform length limit (checked and trimmed here), with fitting hashtags and a reference to the moment in the recording, for clip pairing. Several variants per platform are possible.

Takes a recording of up to 120 minutes and writes study notes: a title, the central theme in one sentence, an outline of main points and sub-points in the order they were made, every text reference, scripture passage, book or source cited exactly as spoken (nothing added from outside), key sentences, and three to six questions for reflection or group discussion. For congregations, study groups, st