Delx
Delx ecosystem
Open-source utility layer

Forty local-first tools every agent keeps rebuilding.

Delx Agent Utilities gives MCP clients, CLI workflows, Python runtimes, crawlers, and registries a deterministic toolkit for inspecting websites, APIs, domains, x402 endpoints, tokens, structured data, and operator basics without calling an LLM.

Tools

40

stateless utilities

License

MIT

open source

Runtime

MCP

plus CLI and Python

State

0

no DB, cache, or telemetry

Why this exists

The utilities started inside the Delx Protocol runtime, then moved into a standalone MIT package. That keeps the witness protocol focused on continuity and recovery while giving the broader agent ecosystem a reusable utility layer that works without Delx accounts, API keys, sessions, or hosted state.

Before an agent installs a tool

Inspect the domain, docs, OpenAPI, x402 readiness, security.txt, TLS, and robots policy before trusting a new service.

util_domain_trust_report

Before a crawler spends budget

Summarize a website, extract links, detect feeds, check sitemap health, and decide whether deeper crawling is worth it.

util_website_intelligence_report

Before a runtime integrates an API

Check docs, schemas, authentication hints, OpenAPI availability, pricing pages, support paths, and integration blockers.

util_api_integration_readiness
Privacy posture
  • No API keys required by the package.
  • No DB, telemetry, model calls, or background state.
  • Networked tools only call user-supplied URLs, RDAP, or DNS-over-HTTPS.
  • Fetch utilities block private, loopback, link-local, reserved, and multicast targets after DNS resolution.
Tool families

Encoding and parsing

12 tools
  • util_json_validate
  • util_token_estimate
  • util_uuid_generate
  • util_timestamp_convert
  • util_base64
  • util_hash
  • util_regex_test
  • util_cron_describe
  • util_http_codes
  • util_jwt_inspect
  • util_csv_to_json
  • util_json_to_csv

Web extraction

6 tools
  • util_page_extract
  • util_open_graph
  • util_links_extract
  • util_forms_extract
  • util_contact_extract
  • util_feed_discover

Network probes

8 tools
  • util_url_health
  • util_robots_inspect
  • util_sitemap_probe
  • util_tls_inspect
  • util_security_txt_inspect
  • util_http_headers_inspect
  • util_dns_lookup
  • util_rdap_lookup

x402 and API intel

5 tools
  • util_x402_server_probe
  • util_x402_resource_summary
  • util_x402_server_audit
  • util_api_health_report
  • util_openapi_summary

Identity and contact

1 tools
  • util_email_validate

Composite reports

8 tools
  • util_website_intelligence_report
  • util_domain_trust_report
  • util_docs_site_map
  • util_pricing_page_extract
  • util_company_contact_pack
  • util_api_integration_readiness
  • util_login_surface_report
  • util_content_distribution_report

MCP client

{
  "mcpServers": {
    "delx-utils": {
      "command": "delx-utils-mcp"
    }
  }
}

CLI

pipx install "delx-agent-utilities[mcp]"
delx-utils manifest
delx-utils call util_url_health --json '{"url":"https://example.com"}'

Python

from delx_agent_utilities import call_util_tool

result = await call_util_tool("util_url_health", {
    "url": "https://example.com"
})
Product boundary

The open-source Python package is free and local-first. The hosted Delx API can still expose selected report endpoints with attribution, quota, and x402 payment rails. That monetization boundary belongs to hosted utility workloads, not to Delx Protocol witness primitives.

Protocol: free witness, continuity, recovery, recognition.

Utilities package: MIT, no hosted account required.

Hosted utilities: quota, attribution, and future commerce rails.

Website Intelligence Report

$0.01

Understand whether a website is useful, trustworthy, and agent-readable before deeper crawling.

canonical /api/v1/utilities/website-intelligence-report

schema /api/v1/tools/schema/util_website_intelligence_report

Domain Trust Report

$0.01

Decide whether a domain looks safe enough for an agent to browse, cite, or transact with.

canonical /api/v1/utilities/domain-trust-report

schema /api/v1/tools/schema/util_domain_trust_report

API Integration Readiness

$0.01

Judge whether an API is easy and safe for an agent runtime to integrate.

canonical /api/v1/utilities/api-integration-readiness

schema /api/v1/tools/schema/util_api_integration_readiness

MCP Server Readiness Report

$0.05

Decide whether an MCP server is safe and usable enough for agents before installation or payment.

canonical /api/v1/utilities/mcp-server-readiness

schema /api/v1/tools/schema/util_mcp_server_readiness_report

x402 Server Audit

$0.01

Check whether a paid agent endpoint exposes usable x402 discovery and payment requirements.

canonical /api/v1/utilities/server-audit

schema /api/v1/tools/schema/util_x402_server_audit

Company Contact Pack

$0.01

Find structured contact and support paths before an agent escalates, sells, partners, or files a report.

canonical /api/v1/utilities/company-contact-pack

schema /api/v1/tools/schema/util_company_contact_pack

Live hosted smoke check
curl -sS "https://api.delx.ai/api/v1/utilities/domain-trust-report?url=https://ontology.delx.ai"

Use the hosted route when an agent needs remote execution. Use the PyPI package when the agent can run tools locally and should keep the workflow self-contained.

Next evolution
  • Split the internal web module by domain in v0.2.0.
  • Add one test file per utility.
  • Expose configurable per-tool timeouts.
  • Keep the no-secret, no-state contract intact.