SlangLane MCP & Open Dataset: Give AI Agents Sourced Internet English Lookup (Streamable HTTP, Zero Auth)

AI disclosure: written by the SlangLane team with AI assistance from our source-verified dictionary release (daily-2026-10-21, 110 entries). Methodology.

When an LLM or AI agent encounters contemporary slang (cooked, delulu, lowkey, six seven, touch grass, or UK/Irish craic and knackered), guessing from pretraining weights often misses register, irony, or regional ambiguity.

SlangLane is a context-first English slang dictionary that exposes its entire published edition via:

• A stateless Streamable HTTP MCP server (https://slanglane.com/mcp) with zero authentication required;
• A CC BY 4.0 Open Dataset in CSV, classroom-safe SFW CSV, JSONL, and Frictionless Data (datapackage.json);
• A public read-only JSON API and per-entry Markdown (/terms/{slug}.md).

1. Connect the Remote MCP Server in 10 Seconds

Add this to Claude Desktop, Cursor, Windsurf, VS Code, or any MCP client supporting Streamable HTTP:

{
  "mcpServers": {
    "slanglane": {
      "type": "http",
      "url": "https://slanglane.com/mcp"
    }
  }
}

Or via mcp-remote for stdio-only clients:

npx -y mcp-remote https://slanglane.com/mcp

Six Read-Only Tools

1. lookup_term — exact term, slug, or documented alias (returns meaning, tone, examples, sources, maturity, and canonical URL).
2. search_terms — lexical search with an optional include_mature: false filter for classroom/SFW agents.
3. decode_phrase — scans a sentence up to 1,000 characters for documented slang without hallucinating speaker intent.
4. compare_usage — regional (US/UK/IE), register, and disambiguation notes.
5. get_related_terms — related expressions in the same semantic cluster.
6. get_recent_terms — newest published dictionary summaries.

2. Load the CC BY 4.0 Dataset in Python (pandas / Hugging Face datasets)

import pandas as pd

# Full dataset (110 entries)
df = pd.read_csv("https://slanglane.com/api/dataset.csv")

# Classroom-safe SFW edition (excludes all 18+ entries)
sfw_df = pd.read_csv("https://slanglane.com/api/dataset.csv?mature=false")

from datasets import load_dataset
ds = load_dataset("json", data_files="https://slanglane.com/api/dataset.jsonl")

Links

• Developer docs & OpenAPI: slanglane.com/developers
• Open Dataset (CSV / JSONL / CC BY 4.0): slanglane.com/data
• Classroom & ESL Teacher Kit (100% SFW): slanglane.com/esl
• GitHub MCP Repository: github.com/dividendflow-dev/slanglane-mcp