mcpField
B
brand-monitor

brand-monitor

github.com/openclaudiaMarketing & growth722447 installsUpdated 1w ago
Repository

About brand-monitor

A brand-monitoring skill for marketing and PR teams that uses the Brand.dev API to find web mentions, assess sentiment, and identify visibility opportunities. It also supports monitor setup and competitor comparisons.

Key features

  1. 01Search brand mentions by date and relevance
  2. 02Retrieve structured brand information
  3. 03Find logos appearing across the web
  4. 04Create and review mention monitors
  5. 05Analyze sentiment and compare competitors

Use cases

  1. 01Reviewing brand coverage over a selected period
  2. 02Finding negative mention spikes for follow-up
  3. 03Checking logo use and media placements
  4. 04Comparing brand visibility with competitors

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/openclaudia/openclaudia-skills --skill brand-monitor

After installing, type/brand-monitorin Claude Code to confirm it loaded.

Details

From the project README

Brand Monitor

Track brand mentions, analyze sentiment, and discover PR opportunities using the Brand.dev API.

Prerequisites

Requires BRANDDEV_API_KEY set in .env, .env.local, or ~/.claude/.env.global.

echo "BRANDDEV_API_KEY is ${BRANDDEV_API_KEY:+set}"

If the key is not set, instruct the user:

You need a Brand.dev API key. Get one at https://brand.dev/ Then add BRANDDEV_API_KEY=your_key to your .env file.

API Base

All requests go to https://api.brand.dev/v1/ with the header Authorization: Bearer {BRANDDEV_API_KEY}.


1. Brand Search

Search for mentions of a brand name across the web.

Endpoint
GET https://api.brand.dev/v1/brand/search
Parameters
ParamTypeDescription
querystringBrand name or phrase to search
limitintNumber of results (default 20, max 100)
offsetintPagination offset
sortstringrelevance or date
from_datestringStart date (YYYY-MM-DD)
to_datestringEnd date (YYYY-MM-DD)
Example curl
curl -s -H "Authorization: Bearer ${BRANDDEV_API_KEY}" \
  "https://api.brand.dev/v1/brand/search?query=YourBrand&limit=20&sort=date"
Response Parsing
curl -s -H "Authorization: Bearer ${BRANDDEV_API_KEY}" \
  "https://api.brand.dev/v1/brand/search?query=YourBrand&limit=20" \
  | python3 -c "
import json, sys
data = json.load(sys.stdin)
for m in data.get('results', []):
    print(f\"Source: {m.get('source','')}  |  Title: {m.get('title','')}  |  Sentiment: {m.get('sentiment','n/a')}  |  Date: {m.get('published_at','')}\")
    print(f\"  URL: {m.get('url','')}\")
    print()
"

2. Brand Info Lookup

Get structured brand information for any company or product.

Endpoint
GET https://api.brand.dev/v1/brand/info
Parameters
ParamTypeDescription
domainstringCompany domain (e.g., stripe.com)
namestringBrand name (alternative to domain)
Example curl
curl -s -H "Authorization: Bearer ${BRANDDEV_API_KEY}" \
  "https://api.brand.dev/v1/brand/info?domain=stripe.com"
Response Fields
  • name: Official brand name
  • domain: Primary domain
  • description: Brand description
  • industry: Industry classification
  • founded: Year founded
  • headquarters: Location
  • social_profiles: Links to social media
  • logos: Brand logo URLs
  • colors: Brand color palette
  • employees_range: Company size estimate

3. Logo Detection

Detect brand logos in images across the web.

Endpoint
GET https://api.brand.dev/v1/logo/search
Parameters
ParamTypeDescription
brandstringBrand name to search for
domainstringFilter to specific domain
limitintNumber of results
Example curl
curl -s -H "Authorization: Bearer ${BRANDDEV_API_KEY}" \
  "https://api.brand.dev/v1/logo/search?brand=YourBrand&limit=20"

Use logo detection to find:

  • Unauthorized logo usage
  • Partner and sponsor visibility
  • Event coverage and media placements
  • Counterfeit product listings

4. Mention Tracking

Set up ongoing tracking for brand mentions.

Create a Monitor
POST https://api.brand.dev/v1/monitors
Body
{
  "name": "My Brand Monitor",
  "keywords": ["YourBrand", "Your Brand", "yourbrand.com"],
  "exclude_keywords": ["unrelated term"],
  "sources": ["news", "blogs", "social", "forums", "reviews"],
  "languages": ["en"],
  "notify_email": "[email protected]"
}
Example curl
curl -s -X POST -H "Authorization: Bearer ${BRANDDEV_API_KEY}" \
  -H "Content-Type: application/json" \
  "https://api.brand.dev/v1/monitors" \
  -d '{
    "name": "Brand Alert",
    "keywords": ["YourBrand"],
    "sources": ["news", "blogs", "social"],
    "languages": ["en"]
  }'
List Monitors
curl -s -H "Authorization: Bearer ${BRANDDEV_API_KEY}" \
  "https://api.brand.dev/v1/monitors"
Get Monitor Results
curl -s -H "Authorization: Bearer ${BRANDDEV_API_KEY}" \
  "https://api.brand.dev/v1/monitors/{monitor_id}/mentions?limit=50&sort=date"

5. Sentiment Analysis

Analyze sentiment of brand mentions.

Endpoint
GET https://api.brand.dev/v1/brand/sentiment
Parameters
ParamTypeDescription
querystringBrand name
from_datestringStart date
to_datestringEnd date
granularitystringday, week, or month
Example curl
curl -s -H "Authorization: Bearer ${BRANDDEV_API_KEY}" \
  "https://api.brand.dev/v1/brand/sentiment?query=YourBrand&from_date=2024-01-01&to_date=2024-03-31&granularity=week"
Sentiment Scores
  • Positive (> 0.3): Praise, recommendations, positive reviews
  • Neutral (-0.3 to 0.3): Factual mentions, news coverage
  • Negative (< -0.3): Complaints, criticism, negative reviews

6. Competitor Mention Comparison

Compare brand mention volume and sentiment against competitors.

Workflow
  1. Search mentions for your brand and each competitor
  2. Compare mention counts over the same time period
  3. Compare sentiment distributions
  4. Identify sources where competitors get mentioned but you do not
# For each brand, get mention counts
for brand in "YourBrand" "Competitor1" "Competitor2"; do
  count=$(curl -s -H "Authorization: Bearer ${BRANDDEV_API_KEY}" \
    "https://api.brand.dev/v1/brand/search?query=${brand}&limit=1" \
    | python3 -c "import json,sys; print(json.load(sys.stdin).get('total',0))")
  echo "${brand}: ${count} mentions"
done

Workflow: Full Brand Audit

When asked for a comprehensive brand monitoring report:

Step 1: Brand Info

Pull structured brand data for context.

Step 2: Mention Volume

Search for brand mentions over the last 30/90 days. Count total mentions and break down by source type.

Step 3: Sentiment Analysis

Get sentiment trends. Flag any negative spikes and investigate root causes.

Step 4: PR Opportunities

From mention data, identify:

  • High-authority sites that mention competitors but not you
  • Journalists who cover your industry
  • Trending topics where your brand could contribute
  • Unanswered questions about your brand on forums
Step 5: Logo/Visual Presence

Search for logo appearances. Flag unauthorized usage.

Step 6: Report

Present findings as:

## Brand Monitoring Report: {Brand}

### Overview
- Total mentions (last 30 days): X
- Sentiment breakdown: X% positive, X% neutral, X% negative
- Top sources: ...

### Sentiment Trend
[Weekly trend data]

### Top Positive Mentions
1. [Source] - [Title] - [URL]
2. ...

### Negative Mentions Requiring Attention
1. [Source] - [Title] - [URL] - [Issue summary]
2. ...

### PR Opportunities
1. [Publication] covers [topic] - pitch angle: ...
2. [Journalist] recently wrote about [topic] - pitch angle: ...

### Competitor Comparison
| Metric | Your Brand | Competitor A | Competitor B |
|--------|-----------|-------------|-------------|
| Mentions | ... | ... | ... |
| Positive % | ... | ... | ... |
| Top Source | ... | ... | ... |

### Action Items
- [ ] Respond to [negative mention]
- [ ] Pitch [publication] about [topic]
- [ ] Update brand listing on [platform]

Error Handling

StatusMeaning
401Invalid or expired API key
403Insufficient permissions for this endpoint
404Resource not found (check monitor ID)
429Rate limit exceeded - wait and retry
500Server error - retry after a few seconds

Tips

  • Use exact brand name + common misspellings as keywords
  • Exclude your own domain to avoid self-mentions
  • Set up monitors for competitor brands too
  • Check mentions weekly at minimum; daily during launches or crises
  • Export negative mentions to a spreadsheet for customer support follow-up

Comments

Markdown supported