ProjectTech4DevAI / ProjectTech4DevAI/kaapi-backend

AI for Content Generation - Shujaaz Video podcast

Offen
#719 2 Kommentare 0 Reaktionen 1 zugewiesene Person Auf GitHub ansehen

Dieses Issue hat noch niemand übernommen.

research sub-parent
Vorherrschende Sprache
Python
Sterne
18
Forks
10
Ø Merge
2 T. 20 Std.
Gemergte PRs (30 T.)
14

Beschreibung

Objective: Use AI to create smart content featuring Shujaaz characters

Problem : Shujaaz currently lack an efficient way to create video podcasts featuring Shujaaz characters. The existing process is cumbersome and has limitations, causing delays and inefficiencies in content production.

Requirement: Use AI to create Video podcast with DJ Boyie and Maria Kim as the speaker. The speaker should be able to reflect the persona of those characters. The communication language will be Sheng (Swahili + English) and the script should follow the free flowing discussion surrounding the current relevant topics for the youths (e.g SRH, Career, Entrepreneurship, Drug abuse etc). Length of video will be 3-4 mins.

Current Users consuming this content: Kenya’s youths (Majorly 18-24 M/F)

Existing Process by Shujaaz team:

  1. Record an actual video (with voice + video + Free flowing discussion on topic) with real users
  2. Upload DJB and Maria Kim images to nano banana to generate the image for podcast studio
  3. Pair the character with the video : Uses Kling to combine 1 and 2 and create animation on top, with output of 15 second content
  4. Stitch 15 seconds video together managing the overlap issue
  5. Post processing or after effect (adding music etc)

Current challenges in existing process:

  1. Kling gives only 15 second video
  2. Stitching issues

Scope for POC
v1: Solve for step 3&4 in existing process - Animation stitching problem

  • Take existing script (3 min around) + voice + image (step 1,2)
  • Solve for step 3 to minimize/eliminate stitching
  • Deliverables: Suggest 1-2 approaches/tool for solving above - list down Tech cost, Manual effort in time, Manual editing ease, other pros n cons. Off the shelf tools are preferred.

v2: Solve for voice generation - accent, energy, tone ?
v3: Solve for script generation - as per topic and guidelines, slang, sheng, flow ?

Other info

  • Kling provides some 1k free credits
  • Characters:
    • DJB: Barasa (18-24 Male)
    • Maria Kim: Nina (18-24 Female)

Resources:

  1. Existing Process by Shujaaz
  2. Character image, video library by Shujaaz
  3. Original detailed PRD by T4D

Beitragsleitfaden

Beitragsleitfaden öffnen

Erste Schritte

  1. Lies das ganze Issue und danach den Beitragsleitfaden des Projekts.
  2. Schreib ins Issue, dass du es übernimmst — das erspart doppelte Arbeit.
  3. Forke das Repository und arbeite in einem Branch.
  4. Öffne einen Pull Request, der die Issue-Nummer nennt.

Bewertung

Dieses Issue wurde noch nicht bewertet.

Neue Issues direkt in Ihr Postfach

Eine kurze Übersicht über anfängerfreundliche GitHub-Issues.