Mindcase for Social Listening Across Platforms

A product mention doesn't show up on one platform. It shows up on Reddit as a question in a niche subreddit, on Twitter/X as a one-line complaint, on TikTok as a review video, and on Instagram as a comment under someone else's post. Checking each platform by hand means four tabs, four search boxes, and four different ideas of what "relevant" means.

If you're an agent (or building one) reading this rather than a human, the full machine-readable schema for every endpoint below lives at mindcase.co/skills.md.

What makes cross-platform social listening hard?

It's not that the data is hard to find — it's that each platform ranks relevance by a different metric, and none of them are directly comparable. A Reddit thread's score is upvotes minus downvotes. A tweet's reach is views. A TikTok video's traction is a mix of views, likes, and shares. An Instagram post's engagement leans on comments as much as likes. Pulling all four and reading them side by side, unranked, just moves the noise problem from four tabs into one feed.

How do you pull mentions across all four platforms in one pass?

Each platform has its own search-style input, so the same agent turn fires four separate calls with the same keyword and merges the results. Here's Reddit Posts and TikTok Posts — Twitter/X Tweets and Instagram Posts take the same keyword-search shape.

curl -X POST "https://api.mindcase.co/v1/data/reddit/posts/run?wait=true" \
-H "Authorization: Bearer $MINDCASE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
  "params": {
    "keyword": "retinol serum",
    "sortBy": "top",
    "timeRange": "week",
    "maxResults": 50
  }
}'

# $0.005 per post returned

curl -X POST "https://api.mindcase.co/v1/data/tiktok/posts/run?wait=true" \
-H "Authorization: Bearer $MINDCASE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
  "params": {
    "searchQueries": "retinol serum",
    "resultsPerPage": 50
  }
}'

# $0.002 per post returned

Four calls, one keyword, four platforms — run in parallel rather than one after another, since none of them depend on each other's output.

How do you tell signal from noise on each platform?

Don't rank the merged results by one shared score — there isn't one. Rank each platform by the metric that platform actually uses to surface relevance, then read the top of each list:

• Reddit — sort by upvotes (or use sortBy: "top" directly in the request, like the example above). A post with 40 upvotes in a niche subreddit is a stronger signal than one with 2, regardless of raw view counts elsewhere.
• TikTok — views and likes together. A video with high views but a low like ratio is often algorithmic reach, not genuine reaction.
• Twitter/X — views for reach, replies for whether people are actually engaging with the claim rather than scrolling past it.
• Instagram — comments weighted more than likes. A comment is a higher-effort signal than a tap, and Instagram's own feed algorithm already leans on it the same way.

This is the same idea this blog covered for a single platform: comment count on Reddit is a better quality filter than raw post volume. The difference here is doing it per-platform, in parallel, instead of assuming one metric works everywhere.

What can't you get from Mindcase?

Two cases.

Private or login-gated content. A locked subreddit, a protected Twitter/X account, a private Instagram profile — none of that is reachable without an authenticated session, and none of these four endpoints have one. That's a public-data boundary, not a gap specific to social listening.

Sentiment scoring. None of the four endpoints return a sentiment label or score — you get the raw post text, author, and engagement numbers. Scoring tone is a step you run yourself, typically by handing the pulled text to an LLM after retrieval, not something baked into the API response.

Which endpoint should you use for which job?

What that costs in practice: pulling 50 results from each of the four endpoints above — 200 posts total across Reddit, Twitter/X, TikTok, and Instagram — comes to $0.25 + $0.0125 + $0.10 + $0.10, or $0.4625 for one full sweep. Twitter/X is priced an order of magnitude below the other three per row, which is why a "check what's being said everywhere" pass skews toward pulling more tweets and fewer, more targeted Reddit and Instagram results if cost is a factor at high frequency.

How do you use social listening data responsibly?

Everything above is public post data — no login, no scraping behind an authentication wall. What matters is what you do with it afterward: don't attribute a handful of loud posts to an entire customer base, keep author identifying details out of anything you publish externally, and check your own legal team's guidance on using social posts as evidence in marketing claims, since that varies by where you operate, not where the poster is.

What's the bottom line?

Four platforms, four different relevance signals, one agent turn. Pulling Reddit, Twitter/X, TikTok, and Instagram together only saves time if you rank each one by what actually predicts relevance on that platform — merging them into one undifferentiated feed just moves the noise problem instead of solving it.

FAQ