PILLAR 05 / DISTRIBUTION AND KITS

Where agents find Caelara, and what you can build on it.

Every suite publishes an MCP server, an agent card and an x402 manifest. The packages below install the same tools into an agent, and the six starter kits are small white-label sites you deploy yourself.

PACKAGES AND REGISTRY

Install and discover

ItemWhere
@caelaralabs/shield-mcpnpm · source
@caelaralabs/mcp (all ten suites)npm · source
@caelaralabs/llms-txt, @caelaralabs/agent-guardllms-txt · agent-guard
elizaOS pluginplugin-caelara-shield
Official MCP Registrycom.caelaralabs/*
Agent cards (A2A)/.well-known/agent.json indexes the ten suite cards; each suite serves its own at /.well-known/agent.json
Guides for agentsllms.txt · llms-full.txt · openapi.json
WHITE-LABEL STARTER KITS

Six small sites you run yourself

AI Search Readiness Auditor

Audits a site for llms.txt, AI crawler access, structured data, SPF, DMARC, headers and certificates. Two composite calls per audit: $0.05 USDC.

ai-seo-auditor-kit · Use this template →

Token and Mint Safety Scanner

A read-only safety screen for EVM tokens and Solana mints. One composite and one atomic call per scan: $0.03 USDC. No wallet connection and no transactions.

onchain-safety-scanner-kit · Use this template →

Prompt Firewall

Screens text for prompt injection, hidden characters, personal data and secrets, with an allow, review or block verdict. One Shield call per screening: $0.015 USDC.

prompt-firewall-proxy-kit · Use this template →

Lead Enrichment

Checks whether a work email can receive mail and looks up the company behind a domain. Up to two calls per lead: $0.055 USDC, or $0.030 for a free-mail address.

b2b-lead-enrichment-kit · Use this template →

Invoice Extractor

Turns the text of an invoice into structured JSON with consistency checks. One composite call per invoice: $0.025 USDC.

invoice-json-extractor-kit · Use this template →

Solana Mint Guard

A read-only Solana mint check: authorities, supply, holder concentration and a safety score. Two calls per check: $0.03 USDC. Includes a JSON endpoint for scripts; no chat bot is included.

solana-mint-guard-bot · Use this template →

All six are MIT licensed Cloudflare Pages projects. Your Caelara key stays in the server-side function and is never sent to the browser. Add your own sign-in and billing before you charge anyone. Prices above are the upstream per-call prices.

WEBSITES, EMAIL AND REAL-TIME CHAT

Protect every place untrusted text gets in or out

Caelara Shield by Caelara Labs and the Routed API Suites by Caelara Labs screen the places where untrusted text and links enter or leave your product. Every check is a heuristic: a clean result is not proof that something is safe.

Websites and web apps

  • Forms and chatbots: screen what visitors type for prompt injection, hidden characters, personal data and secrets POST /v1/shield/prompt $0.015
  • Untrusted URLs: fetch them through the SSRF-guarded proxy, with the page scanned on the way back ANY /v1/proxy/* $0.025
  • Security headers: grade a site’s HTTP security headers GET /v1/security-headers-audit $0.005
  • Trust badges: embed a free, keyless domain-trust or AI-readiness badge benchmarks.caelaralabs.com/verify $0

Email

  • Sender authentication: grade SPF and DMARC for a sending domain GET /v1/spf-dmarc-grader $0.005
  • Links in a message: check the trust of each domain an email links to GET /v1/domain-trust-composite $0.025
  • Outbound leaks: catch keys, credentials and personal data before mail goes out, or mask them POST /v1/shield/output · POST /v1/pii-redactor $0.015 · $0.005

Real-time chat and everyday apps

  • Messages in: screen each message before your app or agent acts on it POST /v1/shield/prompt $0.015
  • Replies out: validate each reply for leaks, exfiltration links and active content POST /v1/shield/output $0.015
  • Inline speed: Shield’s own checks take well under a millisecond in the published benchmark; total time per call depends on the network and on any suite check it runs benchmarks.caelaralabs.com/shield

Audit receipts

  • On every paid screening: a SHA-256 receipt over the surface, verdict, risk level, time and a digest of your request audit_receipt included
  • Independent check: anyone can confirm a receipt, without a key GET /v1/shield/receipt/:hash $0
  • Hash only: no request text is stored; receipts are kept 400 days and are not anchored on a blockchain

Check a message, link or text for free at caelaralabs.com/check/, start with the free trial key ($0, 50 calls, no card), then card or PayPal plans from $9 to $299 a month. Get the free key. Powered by Caelara Labs.

INTEGRATIONS

Connect to Slack, HubSpot, Zapier, n8n & Discord

The platform integrations kit (MIT) screens messages, form submissions and links with Caelara Shield by Caelara Labs, masks personal data before it is stored or synced, and keeps a SHA-256 audit receipt for each paid screening. A check that cannot run returns review, never a silent allow.

Slack

A /shield slash command answers privately with the verdict, a redacted copy and the audit receipt. A channel guard posts a threaded warning when a message is flagged. Requests are verified with your Slack signing secret.

HubSpot

Form submissions are screened field by field and redacted before CRM sync, and the lead’s email domain is trust-checked. Webhooks are verified with the HubSpot v3 signature.

Zapier and Make

Step-by-step recipes with the built-in HTTP step: Gmail, Zendesk or Intercom items go through Caelara Shield and flagged ones are labelled or held.

n8n

Importable workflows for a Gmail inbox guard and a Zendesk or Intercom webhook guard. Your key stays in an n8n credential.

Discord

A /shield command over the Interactions endpoint (Ed25519-verified) and a moderator guard that removes leaked credentials and blocked messages.

Uses POST /v1/shield/prompt with redact, POST /v1/shield/output and GET /v1/domain-trust-composite with one X-Caelara-Key. Heuristic screening will miss attacks. Starter kit platform-integrations-kit (MIT); source on request at ops@caelaralabs.com. Powered by Caelara Labs.

DROP-IN MIDDLEWARE

Web, email and chat guard kit

An MIT starter that screens website forms and chat, outgoing email and real-time messages with one key held on your server. A check that cannot run returns review, never a silent allow.

import { createGuard } from './src/index.mjs'
const guard = createGuard()                                  // reads CAELARA_KEY on your server
app.post('/api/contact', guard.middleware({ fields: ['message'] }), handler)   // 422 on block
const mail = await guard.screenEmail({ fromDomain: 'yourcompany.com', body, links })
const chat = await guard.screenInput(message)                // { verdict, findings, sanitized, receipt }

Starter kit web-email-chat-guard-kit (MIT). Source on request at ops@caelaralabs.com.

DROP-IN PRE-BUY GUARD

Solana pre-buy safety guard for bots and agents

Paste this into a Pump.fun bot, a sniper bot or a Solana agent built on solana-agent-kit or elizaOS, and call it before every buy. The mint audit checks mint and freeze authority; Shield returns an allow, review or block verdict for the mint. Anything but allow with both authorities renounced skips the buy.

// Pre-buy guard for Pump.fun bots, sniper bots and Solana agents. Read-only heuristic screening: never touches keys, wallets or transactions.
const H = { 'X-Caelara-Key': process.env.CAELARA_KEY, 'content-type': 'application/json' } // or pay per call with x402 (Base USDC)
export async function preBuyGuard(mint) {
  const [audit, shield] = await Promise.all([
    fetch(`https://crypto.caelaralabs.com/v1/solana-mint-auditor?mint=${encodeURIComponent(mint)}`, { headers: H }).then((r) => r.json()), // $0.005
    fetch('https://shield.caelaralabs.com/v1/shield/tx', { method: 'POST', headers: H, body: JSON.stringify({ token: mint, chain: 'solana' }) }).then((r) => r.json()), // $0.050
  ])
  const ok = shield.verdict === 'allow' && !audit.mint_authority && !audit.freeze_authority
  return { ok, verdict: shield.verdict, findings: shield.findings ?? [], audit } // usage: if (!(await preBuyGuard(mint)).ok) return // skip the buy
}

From solana-mint-guard-bot (examples/prebuy-guard.mjs, MIT).

CallTerms
GET https://crypto.caelaralabs.com/v1/solana-mint-auditor$0.005 USDC per call
POST https://shield.caelaralabs.com/v1/shield/tx$0.050 USDC per call
AuthX-Caelara-Key header (free 50-call key), or keyless x402 USDC on Base per call
ScopeRead-only heuristic screening. It holds no keys, signs nothing and has no custody of any wallet. A pass is not a guarantee.
ATTRIBUTION

Powered by Caelara Labs

Anything you build on Caelara Labs can carry this badge. It needs no key and no domain, and it is cached for an hour.

Powered by Caelara Labs

Markdown

[![Powered by Caelara Labs](https://benchmarks.caelaralabs.com/verify/powered-by.svg)](https://caelaralabs.com)

HTML

<a href="https://caelaralabs.com"><img src="https://benchmarks.caelaralabs.com/verify/powered-by.svg" alt="Powered by Caelara Labs" height="22"></a>