NLP Tools - Sentiment, NER, Toxicity & Language Detection

fasuizu-br/speech-ai-examples

Connects Claude to five NLP endpoints running on ONNX models: toxicity detection, sentiment analysis, named entity recognition, PII detection with optional redaction, and language identification. All endpoints return in under 50ms and run CPU-only. You get structured JSON responses with labels and confidence scores. Useful when you need to scan user content for moderation flags, extract entities like names and organizations, or redact sensitive information like emails and phone numbers before logging or processing. Part of Brainiall's larger API suite that includes speech and image tools. Requires an API key from their portal, billed per call at fractions of a cent.

tools: 6
transport: HTTP
last updated: Mar 5, 2026

View on GitHub

Install

From registry metadata

$ npx add-mcp https://apim-ai-apis.azure-api.net/mcp/nlp/mcp --name nlp-tools

Run in your terminal. add-mcp asks which agents to add it to (Claude Code, Codex, Cursor and more). Add -g to install it for every project.

Review the command, arguments and environment values before installing. MCP servers run with your local permissions.

Other ways

Claude Code

$ claude mcp add --transport http nlp-tools https://apim-ai-apis.azure-api.net/mcp/nlp/mcp

Run in your terminal. Add --scope user to make it available in every project.

Tools

Verified live · 6 tools

Verified live against the running server on Nov 6, 2026.

analyze_toxicityAnalyze text for toxic content. Returns scores for 6 categories: toxic, severe_toxic, obscene, threat, insult, identity_hate. Each score is 0.0-1.0. BERT-based classifier with sub-15ms latency on GPU. Args: text: Text to analyze for toxicity (hate speech, insults, threats). Re...1 params

Analyze text for toxic content. Returns scores for 6 categories: toxic, severe_toxic, obscene, threat, insult, identity_hate. Each score is 0.0-1.0. BERT-based classifier with sub-15ms latency on GPU. Args: text: Text to analyze for toxicity (hate speech, insults, threats). Re...

Parameters* required

text*string

Text to analyze for toxicity (hate speech, insults, threats)

analyze_sentimentAnalyze text sentiment. Returns positive/negative classification with confidence scores. DistilBERT-based with sub-10ms latency. Multiple domain-specific model variants available. Args: text: Text to analyze for sentiment (positive/negative). model: Model variant -- 'general'...2 params

Analyze text sentiment. Returns positive/negative classification with confidence scores. DistilBERT-based with sub-10ms latency. Multiple domain-specific model variants available. Args: text: Text to analyze for sentiment (positive/negative). model: Model variant -- 'general'...

Parameters* required

modelstring

Model variant: 'general' (default), 'financial', 'twitter'default: general

text*string

Text to analyze for sentiment (positive/negative)

extract_entitiesExtract named entities (NER) from text. Identifies persons, organizations, locations, and miscellaneous entities with span offsets and confidence scores. BERT-NER based with sub-50ms latency. Args: text: Text to extract named entities from. Returns: dict with keys: - entities...1 params

Extract named entities (NER) from text. Identifies persons, organizations, locations, and miscellaneous entities with span offsets and confidence scores. BERT-NER based with sub-50ms latency. Args: text: Text to extract named entities from. Returns: dict with keys: - entities...

Parameters* required

text*string

Text to extract named entities from (persons, organizations, locations)

detect_piiDetect personally identifiable information (PII) in text. Finds emails, phone numbers, SSNs, credit cards, IP addresses, and person names. Optionally returns redacted text with PII replaced by type labels (e.g. [EMAIL], [PHONE]). BERT-NER + regex ensemble. Args: text: Text to...2 params

Detect personally identifiable information (PII) in text. Finds emails, phone numbers, SSNs, credit cards, IP addresses, and person names. Optionally returns redacted text with PII replaced by type labels (e.g. [EMAIL], [PHONE]). BERT-NER + regex ensemble. Args: text: Text to...

Parameters* required

redactboolean

If true, return redacted text with PII replaced by [TYPE]default: false

text*string

Text to scan for personally identifiable information

detect_languageDetect the language of text. Supports 176 languages using fastText. Sub-1ms inference latency. Returns ISO 639-1 codes with confidence scores. Args: text: Text to identify the language of. top_k: Number of top language predictions to return (default: 3). Returns: dict with key...2 params

Detect the language of text. Supports 176 languages using fastText. Sub-1ms inference latency. Returns ISO 639-1 codes with confidence scores. Args: text: Text to identify the language of. top_k: Number of top language predictions to return (default: 3). Returns: dict with key...

Parameters* required

text*string

Text to identify the language of

top_kinteger

Number of top language predictions to returndefault: 3

check_nlp_serviceCheck health status of NLP API services and loaded models. Returns: dict with keys: - status (str): 'healthy' or error state - models (dict): Loaded model status per capability - version (str): API version

Check health status of NLP API services and loaded models. Returns: dict with keys: - status (str): 'healthy' or error state - models (dict): Loaded model status per capability - version (str): API version

No parameters. Call it with no arguments.