Technology Preview

    Integrate Loqate into AI Agents & multi-agent systems using GBG Reach CLI or MCP

    Integrate Loqate into AI Agents & multi-agent systems using our CLI or MCP. It verifies addresses, emails, and phone numbers in real time — returning a confidence score, a policy-aware recommendation, and a full audit trail in one command.

    CLI recommended
    # macOS (Apple Silicon)
    curl -sL https://github.com/gbgplc/lqt/releases/latest/download/lqt_darwin_arm64.tar.gz | tar xz
    sudo mv lqt /usr/local/bin/
    
    export LOQATE_API_KEY=your-key-here
    lqt verify -a "10 Downing St, London, GB" --policy shipping -o json

    CLI is the recommended interface for most use cases — one binary, works in any agent framework, and pipes straight into your tool loop.

    lqt — reach
    $

    Try 'lqt help' to see all commands

    One engine. Two audiences.

    Pipe -o json straight into your agent's tool loop. Switch to -o table when a human needs to eyeball the same call in a terminal. Identical logic, two renderings.

    -o jsonfor agents
    lqt verify ... -o json
    {
      "address": {
        "confidence": 0.95,
        "recommendation": "accept",
        "avc": "V44-I44-P7-100",
        "verification_status": "V",
        "match_level": "premise"
      },
      "email": {
        "confidence": 0.15,
        "recommendation": "reject",
        "response_code": "Valid",
        "risk": "high",
        "flags": [
          "disposable_domain"
        ]
      },
      "phone": {
        "confidence": 0.93,
        "recommendation": "accept",
        "verified": true,
        "number_type": "mobile"
      },
      "overall": {
        "recommendation": "reject",
        "confidence": 0.15
      },
      "policy": "strict"
    }
    -o tablefor humans
    lqt verify ... -o table
    ┌─ Verification ──────────────────────────────────┐
    │                                                 │
    │  Field     Confidence  Recommendation           │
    │  ────────  ──────────  ──────────────           │
    │  Address       0.95    accept    (premise)      │
    │  Email         0.15    reject    (disposable)   │
    │  Phone         0.93    accept    (mobile)       │
    │                                                 │
    │  Overall:      reject  (policy: strict)         │
    └─────────────────────────────────────────────────┘

    Overall = most conservative recommendation. Overall confidence = minimum across fields.

    Capabilities

    Nine tools. Same decisions across CLI, MCP, and Claude Agent Skill.

    Every interface runs the same logic engine. Same input, same confidence score, everywhere.

    verify_address

    Verify an address with a confidence score, AVC-derived match level, and policy-aware recommendation.

    verify_email

    Detect disposable domains, catch-all and fraud-risk flags, and deliverability risk level.

    verify_phone

    Validate phone numbers, detect type, and decide whether to proceed.

    verify_contact

    Run address, email, and phone together. Overall = the most conservative decision.

    parse_address

    Parse and standardize messy input via Claude Haiku — no Loqate credits spent (stdio only).

    list_policies

    List the four built-in policies: strict, shipping, standard, permissive.

    show_policy

    Inspect a policy's thresholds, match-level floor, and reject rules.

    set_policy

    Register a custom policy for the current session (stdio only).

    recommend_policy

    Describe your use case; get a recommended policy (stdio only).

    Confidence scoring

    Deterministic. Numeric. Auditable.

    Every address verification returns a score between 0 and 1 — derived from Loqate's verification signals and country-aware coverage. Same input, same score, every time. Any system can threshold it.

    No letter grades. No opaque interpretations. Just a number your agent — or your workflow — can act on.

    confidence
    0.95
    accept
    V44-I44-P7-100
    Clean US address
    confidence
    0.03
    reject
    U00-I45-P2-100
    Fake address
    confidence
    0.70
    review
    P44-I44-P3-100
    Messy UK address
    confidence
    0.88
    accept
    V44-I44-P4-085
    German address
    Policy-driven decisioning

    The same address. Different answers.

    Strict for KYC. Shipping for delivery. Standard for general verification. Permissive for lead capture. Four built-in policies — or drop in a custom JSON policy tuned to any workflow.

    PolicyAddr confidenceMatch levelEmail confidencePhone requiredUse case
    strict0.90premise0.85yesKYC, fraud prevention, regulated
    shipping0.70street0.50noPhysical delivery, ecommerce
    standard0.55street0.45noGeneral verification (default)
    permissive0.30locality0.30noLead capture, early funnel

    Resolution order: --policy-file --policy LOQATE_POLICY env var → .loqate-policy.json in the working directory → default standard.

    Validate before shipping: lqt policy validate my-policy.json

    {
      "name": "my-custom-policy",
      "description": "Tuned for my use case",
      "address": {
        "min_confidence": 0.65,
        "min_match_level": "street",
        "reject_verification_status": ["U", "R"]
      },
      "email": {
        "min_confidence": 0.50,
        "allow_catch_all": true,
        "reject_disposable": true
      },
      "phone": {
        "min_confidence": 0.40,
        "required": false
      }
    }
    Three interfaces, one engine

    Call it however your agent likes.

    Output parity is a design principle: same input produces the same confidence score across CLI, MCP, and Claude Agent Skill.

    Recommended
    CLI
    best for most

    A single Go binary. Verify, parse, batch-process, pipe the JSON, ship it. Exports a ready-made tool definition for LangChain, OpenAI, or any agent framework.

    lqt verify -a "221B Baker St, London, GB" \
               -e "sherlock@example.com" \
               -p "+442071234567" \
               --policy strict -o json
    MCP server

    Stdio for local agents (9 tools). HTTP for hosted deployments (6 tools) with per-IP rate limiting, request logging, and a /health endpoint. Supports tools & prompts.

    lqt mcp                    # stdio
    lqt mcp --http :8080       # HTTP, rate-limited
    Claude Agent Skill

    Verification inside Claude. Connect once, ask in plain language, get decisions. Install via npx.

    # Claude Code — from GitHub
    npx skills add https://github.com/gbgplc/loqate-skill
    MCP Connector

    Verify contact data in any MCP client.

    Connect Loqate to Claude, Cursor, Windsurf, or any MCP-compatible tool and verify addresses, emails, and phone numbers using plain language.

    Not a fourth interface — this is the HTTP MCP interface, hosted by us so you don't have to deploy it yourself. Just bring your Loqate API key.

    Address verification

    245+ countries & territories

    Email validation

    Deliverability & risk scoring

    Phone checks

    Type, carrier & line status

    Data cleansing

    Parse, correct & standardize

    Step 1 · Connect

    Add the Loqate connector

    Works in any MCP-compatible client — no install required.

    1. 1Open your MCP client's connector settings
    2. 2Add a new connector named "Loqate"
    3. 3Paste the server URL below and save
    Step 2 · Configure your API key

    Provide your Loqate key

    The Loqate tools need an API key to call the service. How you provide it depends on your client.

    Get a free API key
    Claude (claude.ai · Desktop · Code)
    RecommendedOrg / project instructions

    An admin adds your key once — every team member gets it automatically.

    <loqate_api_key>YOUR-KEY</loqate_api_key>
    Settings → Organization → Project instructions
    IndividualUser preferences

    Same tag, scoped to your own profile. Great for personal accounts.

    <loqate_api_key>YOUR-KEY</loqate_api_key>
    Settings → Profile → User preferences
    Quick startSay it in conversation

    Type your key in chat and it's used for the session. Fastest way to try it out.

    Other MCP clients

    Loqate tools accept a key parameter on every call. Most MCP clients let you set default tool parameters in their config — add your API key there so it's passed automatically.

    Try it — just type
    Verify 10 Downing Street, London
    Is test@mailinator.com a real email?
    Check +442071234567
    250+
    Countries verified
    3B
    Annual verifications
    9
    MCP tools
    4
    Built-in policies
    3
    Interfaces
    Get started

    Works with your existing Loqate API keys.

    No new contract. No procurement friction. Install the binary and run your first verification in under five minutes.

    Questions? Email labs@gbg.com

    FAQ

    Frequently asked questions

    How do I install Reach?
    Download the pre-built lqt binary from https://github.com/gbgplc/lqt/releases/latest, place it on your PATH, set LOQATE_API_KEY, and run `lqt verify -a "your address" --policy shipping -o json`. macOS (Apple Silicon) and Windows (x86_64) builds are provided.
    Which agent frameworks does Reach support?
    Anything that speaks MCP (Claude Desktop, Claude Code, Cursor, Cline), the Claude Agent Skill format, or anything that can call a CLI. Use `lqt tool-description` to export a JSON tool schema for LangChain, OpenAI tool-calling, and other frameworks.
    What policies are built in?
    Four built-in policies: strict (KYC), shipping (deliverability), standard (general), and permissive (marketing). Custom policies can be registered per session.
    What does Reach cost?
    The CLI and MCP server are free. Verification calls draw from your existing Loqate account credits. Reach works on a Bring-Your-Own-Key model.