Use case · Arabic podcasts
Produce Arabic podcasts that sound conversational, not synthesised.
Podcasts live and die on the host's voice. Hakim's podcast tier ships catalogue voices in conversational register across 15 Arabic dialects, multi-speaker support for interview formats, and episode-batch rendering that lands an hour-long podcast in under five minutes. Outputs are RSS-feed-ready for Spotify, Apple Podcasts, Anghami, and the regional Arabic podcast platforms.
Three steps from script to RSS
Bring your script
Plain text, with speaker tags for multi-speaker episodes. Hakim accepts up to 50,000 characters per single episode batch (about an hour of audio). Speaker tags follow a
[host]:/[guest]:convention and the renderer assigns the voice you mapped to each role.Map voices to speakers
Pick a host voice from the conversational catalogue (filterable by dialect, gender, age, energy). Add up to five guest voices for interview formats. Or clone the actual host's voice from a 10-second consented sample for shows where the host's identity is the brand.
Render with episode metadata
The batch endpoint produces a single MP3 (or your chosen format) plus an episode metadata block: duration, chapter timestamps if you marked them in the script, ID3 tags pre-filled. Drop straight into your podcast hosting platform (Buzzsprout, Anchor, RSS.com) for distribution to Spotify, Apple Podcasts, and Anghami.
Built for podcast production
Conversational voice register
Most TTS catalogues offer formal news-anchor voices. Podcasts need a different register: warmer, more relaxed, willing to read filler words and incomplete sentences naturally. Hakim's podcast tier was trained on conversational corpora across the Gulf, Egypt, the Levant, and the Maghreb. The voices read like podcast hosts, not news anchors.
Multi-speaker scripts
Interview-format podcasts need at least two voices and often more. Hakim supports up to six distinct voices per episode mapped via speaker tags in the script. The renderer alternates them naturally, including realistic pacing and brief overlap-prevention pauses at speaker transitions.
Host voice cloning for brands
Some podcasts ARE the host's voice. Clone the host from a 10-second consented sample and re-use the cloned voice across every episode. The clone stays consistent over a full season and stays locked to your account so it can't leak to a competing show.
RSS-feed-ready exports
MP3 at 128 kbps for episode files (the standard podcast distribution format), with ID3 tags pre-filled from the episode metadata you provide. Chapter markers honoured in the file (Apple Podcasts and Spotify both surface them in their players). The output is exactly what your hosting platform expects, no transcoding needed.
Request
Multi-speaker episode in five lines
curl https://api.tryhakim.ai/v1/tts/batch \
-H "Authorization: Bearer $HAKIM_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "hakim-arab-v2",
"voice_map": { "host": "layla-egyptian", "guest": "saif-khaleeji" },
"script_url": "https://your-bucket.s3.amazonaws.com/episode-12.txt",
"format": "podcast_mp3",
"webhook_url": "https://your-app.com/hakim/done"
}'Built for these podcast workflows
Solo-host explainer podcasts
Single-narrator shows in the explainer or storytelling format. The host voice carries the entire episode. Voice cloning is the natural pick here so the same voice carries every episode in a season; the conversational catalogue covers the case where the host's voice isn't the brand.
Interview-format podcasts
Two- to six-voice interview shows where the host and guests trade segments. Multi-speaker support handles the voice rotation; speaker tags in the script tell the renderer which voice to use for each line. Realistic transitions and pacing without manual editing.
News and current-affairs podcasts
Daily or weekly news podcasts where the script is written fast and produced fast. Hakim's MSA news-register voices match the convention; episode turnaround under five minutes means the podcast can ship same-day. Anchor and host voices both land in the catalogue.
Branded content and corporate podcasts
Internal corporate podcasts (employee onboarding, executive updates), branded content for marketing, sponsor-funded podcasts where the brand voice is part of the show. Voice cloning lets the brand maintain a single voice presence across every episode without booking a voice actor on every script.
