I want to build an AI agent that generates avatar talking head videos. The avatar should be an attractive blonde female newsreader and she is announcing a new course that allows you to master TIKTOK videos and guides you to your first 1m views, she should be filmed from various angles and have upbeat background music, format to be TIKTOK reels. its a low cost proven formulae where you don't have to show your face and the images are even provided for you. process takes 90 secs a day and if you dont hit 100,000 views we give you your money back further information can be found at https://facelessflipping.com/a/30. The newsreader should present it in a fun semi-serious way and be very enthusiastic
video should be 45 secs long
add
#169 - Create A 5 Second Animated Video From Text Prompt
DESCRIPTION: This skill will return a short 5 second AI-generated animation video based on a given prompt/topic/keyword/product description. The animation will be returned as an MP4 file. This skill can be useful for generating videos that can be added on top of simple text-slide videos to make them more engaging.
INPUT NEEDED: keyword/topic/product description to generate prompt
OUTPUT: URL of MP4 video
#199 - Add Images & Videos On Top Of Existing MP4
DESCRIPTION: This skill will input a collection of images and/or videos (up to 18 in total) and layer them on top of an existing MP4 URL. It will intelligently review the new media and decide where to insert it. For added control, it is recommended that a description of the images/videos, and transcription of the video is provided.
INPUT NEEDED: List of images/videos, ideally with descriptions, and transcription of video
OUTPUT: URL of mp4 video
#222 - Make Image (Especially With Text)
DESCRIPTION: This skill uses an Image AI which is very good at both prompt adherence and generating text. You should use this for any image task which requires text. It can generate images in a range of aspect ratios.
INPUT NEEDED: Information about the information required (the AI will generate its own particular prompt, so the more information inputted the better).
OUTPUT: PNG image URL
#223 - Powerful LLM Prompt-to-Text Response
DESCRIPTION: This skill uses a very powerful LLM to answer any query. It will return a text response. The LLM can only input text and is not multimodal.
INPUT NEEDED: Prompt and any other information required to answer the query.
OUTPUT: Text response from LLM
#190 - Write or rewrite text based on instructions
DESCRIPTION: This task uses powerful LLMs to write text from scratch, using its inherent training knowledge. This is very useful skill, since it is so quick and easy to do, requires no external research data (which can be inconsistent), and since the LLM is so good at following specific prompt instructions
INPUT NEEDED: Detailed text prompt
OUTPUT: Text response
#207 - Get Transcription From MP4 Video URL
DESCRIPTION: This skill will input an MP4 URL and generate a transcription of the audio file, along with exact timings.
INPUT NEEDED: MP4 URL
OUTPUT: Perfect transcription of text with timings separated by line (approx. 3-8 seconds per timing)
#200 - Generate Text Slide Video From MP3 & Transcription (max 2000 characters)
DESCRIPTION: This skill requires an MP3 file and script input, and will then generate a text slide video (in 16:9 traditional YouTube aspect ratio).
INPUT NEEDED: MP3 URL + Transcription (both are required, and transcription requires timings)
OUTPUT: MP4 video URL
#215 - Summarise and Return All Final Output Files
DESCRIPTION: This skill returns a summary of various previous skills that have completed, and is particularly useful when appearing at the end of an entire Agent task, silo or sequence of skills.
INPUT NEEDED: at least one output (perhaps other supplementary ones that may be helpful)
OUTPUT: nice text summary containing 1 or more outputs video in portrait (9:16)
2. just one version of video
3. include captions/subtitles in the video
4. Should specific branding elements (colors, fonts, logos) be incorporated-use colours and logos as found on dotcomecom.com
5. repeat the USps-only 90secs a day, you are guaranteed 100,000 views min on your video or money back, images supplied daily to you, no need to show your face, many members now seeing over 1m views
6. Show the URL at the end with a big call to action to go there NOW
7. background music should be upbeat
8. avatar's expressions be more casual/friendly/flirty
1. The website URL to be used in call to action: facelessflipping.com/a/30
2. Key selling points (already mentioned: low cost, proven formula, no face showing, images provided, 90 seconds daily, 100k views guarantee)
3. Price point less than a few coffees
5.relay that many members have seen the 1m views and one has just hit 5m
I would suggest segmenting this into three main silos:
1. Content Development Silo:
- Script writing for the news anchor
- Background music selection
- Visual storyboarding (planning the camera angles and transitions)
2. Visual Asset Creation Silo:
- AI avatar generation in different angles
- Text overlay design
- Any supporting graphics or animations
- Transition effects
3. Post-Production Silo:
- Video composition
- Audio synchronization
- Format optimization for TikTok
- Caption generation
- Final quality check and platform-specific optimization
These silos would help ensure each component gets proper attention and allows for easier revisions if needed.
1. newsreader avatar to have a white blouse on showing a good cleavage, blond hair shoulder length, blue eyes no rings but settle necklace
2.newsreader should be professional/flirty and excited
3. What is the key message or call to action you want to convey in this video?--ITS JUST 90 secs a day
4. How many angles would you like us to cover for the newsreader? 4
7. For the 5-second animated video, is there a specific scene or message you want to convey? The call to action and the main bullet points
the presentation should follow what is it, what does it do, what do I have to do, how do I get it