Isobanuramiterere
Byemewe mu bucuruzi
moderi 579
Nta kimenyetso cy'amazi
Nta kwiyandikisha gukenewe
Model:
+ GPT-5, Claude, Gemini
Generate full YouTube video metadata: 5 title variants under 60 chars (the cap before truncation in search), a 2,000+ char SEO-optimized description with hook in the first 2 lines (before 'Show More' cut), 30 tags mixed broad-to-specific, 3-5 hashtags, 3 thumbnail-text options, a pinned-comment suggestion, plus a hook script for the first 8 seconds of the video. Tuned for YouTube's 2024 ranking signals: keyword-front-loaded titles + AVD-friendly chaptering.
Ibyatanzwe
Amahitamo yisumbuyeho
Igisubizo
Tokens ziri gushira.
Bona tokens nyinshi
❤️ Ukunda Free.ai? Bwira inshuti zawe!
Iyandikishe kugira ngo ubone umurongo wo gutumira ubone tokens 30,000 kuri buri nshuti.
Urashaka ibindi?
Iyandikishe ku buntu: tokens 30,000/umunsi
Iyandikishe ku buntu
Turimo gukora icyo wasabye...
Kora ibisobanuro by'amashusho ya YouTube ukoresheje AI y'ubuntu. Byateguwe kuri SEO kugira ngo bibonwe cyane.
Uko bikoreshwa Isobanuramiterere
1
Shyiramo ibyo ushaka
Andika inyandiko, ohereza dosiye cyangwa usobanure icyo ushaka. Nta konti ikenewe.
2
Kanda generate
AI yacu ikora icyo wasabye mu masegonda ikoresheje moderi nziza za open-source.
3
Kuramo & usangize
Kuramo, koporora cyangwa usangize igisubizo cyawe. Ni ubuntu ku bikorwa bwite n'iby'ubucuruzi.
Koresha iki gikoresho ukoresheje API
Koresha iki gikoresho mu buryo bwikora uri muri code yawe. REST endpoint ihuye na OpenAI, kwinjira hakoreshejwe Bearer token, nta SDK yiyongereye ikenewe. Igiciro cya tokens ni kimwe n'icyo ku rubuga.
curl -X POST https://api.free.ai/v1/chat/ \
-H "Authorization: Bearer sk-free-..." \
-H "Content-Type: application/json" \
-d '{"model": "qwen7b", "messages": [{"role": "user", "content": "Write an Instagram caption for a product launch."}]}'Ibindi bikoresho bya AI by'ubuntu
Isobanuramiterere - FAQ
Generates full YouTube metadata for 9 content types (long-form 8-30min, Shorts under 60s, tutorial / how-to, review / unboxing, podcast / interview, vlog / lifestyle, channel About / banner, community post, playlist description). Outputs 5 title variants under 60 chars, 2,000-3,000-char SEO description with hook in first 2 lines (before 'Show More' cut at ~150 chars), 30 mixed-volume tags, 3-5 hashtags, 3 thumbnail-text options, pinned-comment suggestion, and an 8-second hook script for the video opening.
TubeBuddy is $4.50-49/mo, vidIQ $7.50-79/mo, Morningfame $12-39/mo. All three give keyword research + score + competitive analysis but DON'T write the description copy. ChatGPT (free + $20/mo) writes generic descriptions without YouTube-specific structure (chapters, time-stamps, the first-2-line hook rule, banned-clickbait list). Free.ai is purpose-built - content-type-aware system prompts, 9 distinct prompts (Shorts ≠ long-form ≠ podcast). Free.
Yes - Qwen 3 30B on our GPUs. A long-form metadata package (titles + description + tags + thumbnail + hook + pinned) costs about 1,200 tokens. The 30K/day daily pool covers ~8 sessions. Premium models (Claude Sonnet, GPT-5, DeepSeek R1) write punchier hooks and tighter chapter summaries - worth the swap when you're competing in a crowded niche.
(1) Long-form (8-30 min) - full metadata pack, AVD-driven. (2) Shorts (under 60s) - short title + tiny description + #Shorts + 3-sec hook only. (3) Tutorial - problem-first SEO + chapters + tools/links. (4) Review / unboxing - product name in title + verdict cue + affiliate disclosure placeholder. (5) Podcast / interview - guest name first (their audience subscribes). (6) Vlog - vibe-driven, lighter SEO, music credits. (7) Channel About / banner. (8) Community post - under 600 chars, poll-friendly. (9) Playlist description - what unifies the videos.
Yes - system prompt enforces the 60-char cap with [XX/60 chars] badges next to each variant. Sweet spot per VidIQ + TubeBuddy data is 45-55 chars (longer = truncated in mobile search; shorter = looks sparse). The model front-loads the primary keyword for left-of-screen visibility.
YouTube cuts your description at ~150 chars before the 'Show More' button on mobile. Lines 1-2 must hook the viewer + signal what they will get - they're the only thing 80% of mobile users see. The system prompt enforces this hook structure. Below the cut: full description with chapters, links, about / subscribe.
The first 8 seconds of any YouTube video decide retention - 30% of viewers leave by second 8 if the hook is weak (per YouTube Studio data). The hook script is a word-for-word opening: pattern interrupt + state the value + tease what they'll miss if they leave. NEVER 'Hey guys, welcome back to my channel' - that's the #1 viewer-loss line.
No - banned phrases include 'You WON'T BELIEVE', 'GONE WRONG (POLICE CALLED)', '*EMOTIONAL*', 'MUST WATCH', all-caps spam, excessive emoji. YouTube's algorithm deboosts overpromise clickbait per the 2024 trust update - using these tanks your CTR and watch-time long-term.
Yes but less than they were. Per YouTube's official 2023 statement, tags are mostly used to correct misspellings of your title keyword. The system prompt generates 30 mixed-volume tags (broad / medium / long-tail) anyway because they take 5 seconds to add and have small upside. For Shorts, only 15 tags - Shorts uses fewer signals.
No - first session free. Sign up for the 30K/day pool, generation history, premium models. ~8 metadata packages per signup grant.
Yes - 9 explicit language presets (English, Spanish, Portuguese, French, German, Japanese, Hindi, Indonesian, Arabic) with regionally appropriate keyword formatting. The 6 biggest YouTube markets outside English are Spanish, Hindi, Portuguese, Indonesian, Japanese, Arabic - picking your real language drastically outperforms English-everywhere.
Yes - POST to /v1/chat/ with the YouTube metadata system prompt (in template source). Useful for video-management SaaS, agency client deliverables, or video-editor plugins auto-generating metadata. Bearer auth, monthly limits. Docs at /api/.
Wakwemeza iki kigero iki gikoresho?