keyapi-tiktok-influencer-discovery（keyapi 抖音影响者发现）

keyapi-tiktok-influencer-discovery

Discover, profile, and deeply analyze TikTok influencers — from keyword-based search to multi-dimensional performance intelligence.

This skill powers end-to-end TikTok influencer research using the KeyAPI MCP service. It enables you to find creators by keyword or region, retrieve their profile and performance metrics, analyze historical growth trajectories, and benchmark them against ranking data — all through a single, orchestrated workflow.

Use this skill when you need to:

• Identify high-performing influencers for brand collaborations or affiliate campaigns
• Audit a creator's follower growth, engagement rate, and live-stream GMV history
• Build ranked shortlists and compare multiple creators across key performance dimensions
• Track historical trends for competitive intelligence and market positioning

author: KeyAPI license: MIT repository: https://github.com/EchoSell/keyapi-skills

Prerequisites

author: KeyAPI license: MIT repository: https://github.com/EchoSell/keyapi-skills

MCP Server Configuration

All tool calls in this skill target the KeyAPI MCP server:

Server URL : https://mcp.keyapi.ai
Auth Header: Authorization: Bearer $KEYAPI_TOKEN

Setup (one-time):

# 1. Install dependencies
npm install

# 2. Set your API token (get one free at https://keyapi.ai/)
export KEYAPI_TOKEN=your_token_here

# 3. List all available tools to verify the connection
node scripts/run.js --list-tools

author: KeyAPI license: MIT repository: https://github.com/EchoSell/keyapi-skills

Analysis Scenarios

Select one or more nodes based on the research objective. Multiple nodes can be combined for cross-dimensional analysis.

author: KeyAPI license: MIT repository: https://github.com/EchoSell/keyapi-skills

Workflow

Step 1 — Identify Analysis Targets and Select Nodes

Clarify the user's objective and map it to one or more nodes from the table above. Typical entry points:

• Keyword discovery: Start with search_influencers, then optionally deepen with influencer_list_analytics for richer filtering.
• Direct profile lookup: Use get_influencer_detail with a known unique_id (@handle).
• Performance deep-dive: Combine influencer_detail_analytics + influencer_trends_analytics + influencer_videos_analytics.
• Live commerce evaluation: Use influencer_livestreams_analytics + influencer_products_analytics.
• Competitive ranking: Use influencer_ranking_analytics with appropriate category/region filters.

⚠️ Critical: Resolving user_id vs. unique_id Two distinct identifier types are used across endpoints: - unique_id — the user's public @handle (e.g., charlidamelio). User-visible, mutable. - user_id — TikTok's permanent, immutable numeric UID assigned to each account. When a workflow requires nodes that accept different identifier types, always call get_influencer_detail first using the unique_id to obtain both identifiers before proceeding.

Step 2 — Retrieve API Schema

Before calling any node, inspect its input schema to confirm required parameters, data types, and valid enumeration values:

node scripts/run.js --schema <tool_name>

# Example
node scripts/run.js --schema influencer_list_analytics

For analytics nodes, pay particular attention to filter parameters (region, category, date range, follower range, etc.) and confirm the expected page_num/page_size fields.

Step 3 — Call APIs and Cache Results Locally

Execute the required tool calls and persist all responses to the local cache to enable result reuse across sessions and avoid redundant API calls.

Calling a tool (using scripts/run.js):

# Single page call — result is cached automatically
node scripts/run.js --tool <tool_name> --params '<json_args>' --pretty

# Fetch all pages at once (auto-pagination)
node scripts/run.js --tool <tool_name> --params '<json_args>' --all-pages --page-size 50

# Force a fresh call, skip cache
node scripts/run.js --tool <tool_name> --params '<json_args>' --no-cache

Example — search influencers:

node scripts/run.js --tool search_influencers \
  --params '{"keyword":"fitness","region":"US"}' --pretty

Example — filter influencers with analytics (all pages):

node scripts/run.js --tool influencer_list_analytics \
  --params '{"region":"US","influencer_category_name":"Fitness"}' --all-pages

Example — get influencer's latest videos (cursor-based):

# First page: offset=0
node scripts/run.js --tool get_influencer_videos \
  --params '{"unique_id":"charlidamelio","offset":"0"}' --pretty
# Next page: use max_cursor value from previous response as offset

Pagination for analytics endpoints:

All *_analytics endpoints use page_num (1-indexed) and page_size (max 10). run.js injects these automatically if not specified. Use --all-pages to let run.js iterate all pages and merge the results.

--page-num 1  --page-size 10   → first page (default)
--all-pages                    → all pages merged into one result

Note: get_influencer_videos, get_influencer_followers, get_influencer_following use cursor-based pagination via an offset parameter — not page_num/page_size. Pass "offset":"0" to start, then use the max_cursor (or min_time) value from the response as the next offset.

Cache directory structure:

.keyapi-cache/
└── influencers/
    └── {unique_id}/
        ├── detail.json                  # get_influencer_detail
        ├── analytics.json               # influencer_detail_analytics
        ├── trends.json                  # influencer_trends_analytics
        ├── videos_analytics.json        # influencer_videos_analytics
        ├── livestreams_analytics.json   # influencer_livestreams_analytics
        ├── products_analytics.json      # influencer_products_analytics
        ├── latest_videos.json           # get_influencer_videos
        ├── followers.json               # get_influencer_followers
        ├── following.json               # get_influencer_following
        ├── region.json                  # get_influencer_region
        ├── qr_code.json                 # get_influencer_qr_code
        └── milestones.json              # get_influencer_milestones
└── searches/
    └── influencers/
        └── {md5_of_query_params}.json   # search_influencers, influencer_list_analytics
└── rankings/
    └── influencers_{params_hash}.json   # influencer_ranking_analytics

Cache-first policy:

Before every API call, check whether a cached result already exists for the given entity and node. If a valid cache file exists, load from disk and skip the API call.

Cover image processing:

After each API call, scan all response image URLs. If any URL's host matches echosell-images.tos-ap-southeast-1.volces.com, collect those URLs and call batch_download_cover_images in a single batch request. Replace the original URLs in your working dataset with the converted URLs returned by this node.

Step 4 — Synthesize and Report Findings

After collecting all API responses (from cache or live calls), produce a structured research report:

1. Creator Profile Summary — Name, @handle, follower count, engagement rate, primary niche, and operating region.
2. Performance Analysis — Follower growth curve, average video views, engagement benchmarks, and live-stream GMV history.
3. Content Strategy Insights — Top-performing video themes, posting cadence, product promotion patterns, and audience interaction quality.
4. Competitive Positioning — Ranking within category/region, peer comparisons when analyzing multiple creators.
5. Actionable Recommendations — Best fit use cases (brand sponsorship, affiliate, live commerce), audience-campaign alignment, risk signals (follower authenticity, trend consistency).

Cross-reference multiple data sources where available — for example, correlate influencer_trends_analytics with influencer_livestreams_analytics to identify whether GMV peaks align with follower growth events.

author: KeyAPI license: MIT repository: https://github.com/EchoSell/keyapi-skills

Common Rules

author: KeyAPI license: MIT repository: https://github.com/EchoSell/keyapi-skills

Error Handling