We Tested Every Way to Use Claude Code For Social Media

What are the 12 core subtopics within [PASTE YOUR INPUT HERE] that a content strategy should cover?

Using the SerpAPI, search Google for " [PASTE YOUR INPUT HERE] " and pull the "People Also Ask" questions. Run multiple searches with different related queries (e.g., " [PASTE YOUR INPUT HERE] tips", "how to [PASTE YOUR INPUT HERE] ", " [PASTE YOUR INPUT HERE] for beginners", " [PASTE YOUR INPUT HERE] guide") to collect at least 30 unique "People Also Ask" questions. Remove any repeated questions and return a final numbered list of the top 30.

Use Apify TikTok scraper to find this account's [PASTE YOUR INPUT HERE] top 20 most-viewed posts and capture the opening hook or first line of each video using the transcription from the video. Return the results in a table with columns Engagement Rate, Hook/First Line and Hook Format. Summarize the most common hook formats and identify which formats perform best based on engagement rate using views, likes, and comments. Save all full verbatim transcripts in a single file called transcripts.txt, separated by the video rank number and title so each transcript is clearly labeled.

I have two datasets from my research:

Dataset 1 - Top-performing hook formats from a creator in my niche, ranked by engagement rate:

Dataset 2 - What people are actively searching for (Google "People Also Ask" questions):

Your task:

For each search question, generate a short-form video hook that answers the search intent but is rewritten using one of the top-performing hook formats from Dataset 1. Match each question to whichever hook format would work best for that topic.

Return a table with columns:

Search Question (the original PAA question)
Video Hook (the rewritten hook using a proven format)
Hook Format Used (which format from Dataset 1)
Why This Pairing Works (one sentence explaining why this format fits this topic)

Then pick your top 5 video ideas that have the best chance of going viral based on the combination of high search demand + high-performing format, and explain why.

Using Apify TikTok scraper, scrape the 50 most recent posts from [PASTE YOUR INPUT HERE] . Find the 3 videos most similar in topic to the video below. Transcribe those 3 videos in full - every word spoken, verbatim.

Then, using only those 3 transcripts as your style reference, write a 60 second short-form video script that matches this creator's voice, pacing, vocabulary, tone, and structure.

Topic:

Hook:

Hook Format:

Search Intent:

RULES:
- Write it as a SPOKEN script, not a blog post. Include natural speech patterns from the transcripts.
- Keep sentences short. This is spoken to camera, not read.
- Do NOT use phrases the creator never uses. Stay in their voice.
- Do NOT add stage directions, B-roll suggestions, or visual cues.
- Output ONLY the script, nothing else.

You are a short-form video editor. I'm giving you a script for a 60 second talking head TikTok. Most of the video is just me on camera. Your job is to identify 3 moments in the script where a motion graphic would make the point land harder, then produce a two-column table - left column is the script broken line by line, right column is what the viewer sees. Use "Talking head" for every line that doesn't need a graphic. For the 3 graphic moments, describe exactly what the animation should look like, how long it runs, and whether it overlays on me (green screen) or takes over full screen. Keep descriptions specific enough that a Remotion developer could build them without asking follow-up questions. Here is the script:

Edit this video:

Script - only keep takes that match these lines:

Instructions:

Transcribe the full video using Whisper (small model or better) with word_timestamps=True. Also run ffmpeg silencedetect (noise=-30dB, d=0.3) - Whisper alone can merge separate takes into one segment, so you need silence boundaries to tell them apart.

Identify each take by cross-referencing word timestamps with silence gaps. When the same line appears multiple times, the later take is usually the keeper - the speaker restarted because the earlier one had a stumble or flub. Never splice words from two different takes together.

Remove bad takes - any lines that don't match the script, and any earlier attempts of a line where a cleaner retake exists.

Remove silences - trim dead air between lines, but leave ~0.15s of the original room tone at each cut point so it sounds natural, not jarring.

Trim tightly at the head and tail - start the video right when the first word begins (no leading silence or closed-eyes frames). End each word fully - use the Whisper word end timestamp (not the silence detect start), since silence detect can trigger before a word's final consonant finishes.

Export the final cut as [PASTE YOUR INPUT HERE] .

You are a Remotion developer. I'm giving you a two-column table from a short-form video script. The left column is the script, the right column describes what the viewer sees. Rows marked "Talking head" need nothing from you - skip them. Every other row describes a motion graphic you need to build.

For each motion graphic in the table:
- Build it as a standalone Remotion composition at 1080x1920, 30fps
- If the visual description says "overlay" or "green screen," render with a chroma-key green background (#00FF00) so it can be keyed out in a video editor. Do not use a transparent background - MP4 does not support alpha channels.
- If it says "full screen," use a dark background
- Match the timing to how long it would take to speak the script in that row at a natural pace
- Use clean, modern sans-serif type (Inter or similar), smooth easing, and bold accent colors

Here is my table:

Transcribe the audio from [PASTE YOUR INPUT HERE] using Whisper with word timestamps. Then burn captions onto the video with these specs:

Font: Helvetica Neue Bold (/System/Library/Fonts/HelveticaNeue.ttc, index 1)
Font size: 52px
White text with solid black outline (8px thick)
Sentence case (capitalize at start of new sentences and proper nouns)
Max 5 words per caption
Each caption stays on screen until the next one appears (no gaps)
Centered, positioned at 80% from top of frame
Maintain the original video's resolution and frame rate
Use Pillow for text rendering and ffmpeg overlay filter for compositing
Save output to Desktop with _captioned appended to the filename
Correct any obvious transcription errors based on context

Set up Post Bridge as an MCP server so I can schedule social media posts. Here's my API key: [PASTE YOUR INPUT HERE] . Here are the setup docs: https://www.post-bridge.com/mcp/docs

Schedule this [PASTE YOUR INPUT HERE] to my [PASTE YOUR INPUT HERE] account for [PASTE YOUR INPUT HERE] . Use Post Bridge.

I'm attaching my TikTok analytics exports from TikTok's "Export data" feature. I have the last 365 days of data across 4 categories: Overview, Content, Viewers, and Followers.

Analyze all of this data and give me a complete TikTok performance breakdown using the sections below. Use the 365-day dataset to derive the last 7 days, last 28 days, and last 60 days yourself. Be direct, specific, and reference actual numbers from my data. No fluff - just insights I can act on.

SECTION 1: This Week's Snapshot
Total video views, likes, shares, and profile views for the week
Daily average for each metric
Best and worst performing day this week (and what might explain it)
Net follower change this week (are we growing, flat, or losing?)

SECTION 2: Trend Comparison
This week's daily averages vs. last 28 days vs. last 60 days vs. last 365 days
Put each metric in a simple table so I can see the trend at a glance
Call out any metrics that are clearly trending UP or DOWN
Flag anything that looks like a spike or dip anomaly in the historical data
Follower growth rate: compare the pace of follower change across all time ranges

SECTION 3: Content Analysis
Which videos are still driving the most views right now?
What do my top videos have in common?
What content formats or topics appear in my top performers vs. the ones that underperform?
Which videos have unusually high shares relative to views?
Which videos have unusually high likes-to-views ratios?

SECTION 4: Audience Breakdown
Gender split and what it means for my content strategy
Top countries/regions and whether I should lean into or diversify
New vs. returning viewer ratio
How has the new vs. returning viewer mix changed across time ranges?
Peak active follower hours - when exactly should I be posting?

SECTION 5: What to Double Down On
The content themes/formats that are working (with evidence from the data)
The posting patterns that correlate with better performance
Anything my audience is clearly responding to that I should do MORE of

SECTION 6: What to Fix or Change
Any metrics that are declining and what might be causing it
Content gaps - what's missing based on what performs well?
If I'm losing followers, what does the data suggest about why?
Engagement rate concerns (likes, comments, shares relative to views)

SECTION 7: The Plan
This week: 3 specific things I should do in my next few videos
This month: 2-3 strategic adjustments to test
Ongoing: 1-2 habits to build into my content routine based on what the data says works

I'm going to paste my TikTok analytics report below. It includes recommendations for my account.

Read the recommendations and pick the top 3 that would benefit most from researching what other creators are already doing successfully on TikTok.

For each of those 3, use the Apify TikTok scraper to find top-performing videos from other creators that are executing on that recommendation. I want to see what's actually working right now, not theory.

Based on what you find, give me a 2-week content calendar with specific video ideas that combine my proven strengths from the analysis with what's working for others right now. Include the reference TikTok URLs so I can study them.

--- PASTE YOUR TIKTOK ANALYSIS BELOW THIS LINE ---

Take the TikTok analytics breakdown above and turn it into a single-page, self-contained HTML website I can open in a browser. Requirements:

1. Dashboard-style layout with a clean, modern design (dark mode preferred). Use only inline CSS and vanilla JavaScript - no external dependencies or CDN links.

2. Section for each part of the analysis - make the sections visually distinct with cards or panels. Include all the key numbers, tables, and insights from the breakdown.

3. Charts/visualizations using HTML5 Canvas or inline SVGs for:
   - Daily views over time (line chart)
   - Monthly trend comparison (bar chart)
   - Day-of-week performance (bar chart)
   - Follower growth/loss over time (line chart)
   - New vs. returning viewers (donut or pie chart)
   - Audience geography (horizontal bar chart)
   - Top videos ranked by views (horizontal bar chart)

4. Use the actual numbers from my data - don't use placeholder or dummy data.

5. Make it responsive so it looks good on both desktop and mobile.

6. Include a summary header at the top with the account name, date range, and 3-4 key headline stats as large KPI cards.

7. Color-code trends - green for metrics trending up, red for metrics trending down, yellow/amber for flat or mixed signals.

8. Include the action plan at the bottom as a styled checklist or roadmap section.

I have a short-form video script below. Repurpose it into 3 formats:

1. LinkedIn Post
Written in first person, conversational tone
Open with a strong hook line (not the video hook - write one native to LinkedIn's scroll pattern)
Break into short paragraphs (1-2 sentences each) for mobile readability
End with a clear CTA or question to drive comments
Keep it under 1,300 characters so it doesn't get cut off behind "see more"
No hashtags in the body - add 3-5 relevant hashtags at the very end

2. Email (newsletter style)
Subject line that creates curiosity or promises a specific takeaway
Preview text (the first line someone sees in their inbox) that complements the subject line
Conversational tone, written like I'm talking to one person
Pull the key insight from the script and expand on it slightly - add context that makes sense without watching the video
Include a link placeholder [VIDEO LINK] where I'd drop the video
End with a single CTA - reply, watch, or click

3. Tweet / X Post
Under 280 characters
Punchy, standalone insight pulled from the script
Should make sense without any context
No hashtags, no emojis unless they add meaning
Write 3 variations so I can pick the best one

Rules:
Don't just summarize the video - each format should feel native to its platform
Don't reuse the same hook across formats
Match my voice from the script - don't make it sound more formal or corporate

Here's the script:

Take everything you just did in this conversation and turn it into a skill I can use again. I want to be able to run it anytime by typing a slash command like /skill-name. Set it up so I can plug in different information next time instead of it being locked to what we just worked on.

I have a skill called [PASTE YOUR INPUT HERE] .

Set up a scheduled task that runs it every [PASTE YOUR INPUT HERE] .

It should use the files in [PASTE YOUR INPUT HERE] .

When it's done, [PASTE YOUR INPUT HERE] .