Agent-native

Connect your AI agent to the evidence-backed skill graph

The public side of InTransparency as tools your agent can call: student portfolios their owners made public, company profiles, active job postings, Italian schools and universities, credential checks, ESCO and the algorithm registry. Works with Claude Code, Claude Desktop, Cursor, Zed, ChatGPT, LangChain, LlamaIndex and any OpenAPI-aware host.

Read-only. The tools, and the public endpoints behind them, need no key and no account. No agent can apply for a job or publish one: an agent key only adds watchlists and webhooks, for the account that created it.

Option 1 — MCP server (Claude Code, Claude Desktop, Cursor, Zed)

The Model Context Protocol server gives Claude Code, Claude Desktop and compatible clients tools to search public student portfolios, companies, job postings, and Italian schools and universities, to verify a credential, to map a skill to ESCO and to read the algorithm registry and the other reference data. Every tool is listed below.

Fastest — remote server (one line, no download)

Add the hosted MCP server. Nothing to download, no path to configure:

claude mcp add --transport http intransparency https://www.in-transparency.com/api/mcp

Works in Claude Code, Claude Desktop, Cursor, Zed, and any host that supports remote HTTP MCP servers.

Manual install (stdio fallback)

1. Download the MCP server
Save this file anywhere on your machine:
```
curl -O https://www.in-transparency.com/mcp-server.js
```
2. Register it in Claude Desktop
Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows) and add:
```
{
  "mcpServers": {
    "intransparency": {
      "command": "node",
      "args": ["/ABSOLUTE/PATH/TO/mcp-server.js"]
    }
  }
}
```
Replace /ABSOLUTE/PATH/TO with the folder you saved the file in. Claude Desktop does not start the server from that folder, so a relative path is not found. On Windows, forward slashes work too, as in C:/Users/name/Downloads.
Restart Claude Desktop. A new "intransparency" tool set will appear.

Available tools

Each tool, with the first sentence of the description the server sends to your agent.

intransparency_search_talent: Search the directory of PUBLIC student portfolios.
intransparency_search_companies: Search the InTransparency directory of published company profiles.
intransparency_lookup_company: Fetch a single company profile by slug.
intransparency_search_jobs: Search active job postings.
intransparency_search_schools: Search the Italian school reference: MUR-recognised universities, ITS Academies, and every secondary school (licei, istituti tecnici and professionali, statali and paritarie) from the Ministry's open registry, one row per institute.
intransparency_lookup_job: Fetch a single job posting by ID with full description.
intransparency_verify_credential: Cryptographically verify a W3C Verifiable Credential issued by an InTransparency institution.
intransparency_resolve_esco: Map a free-text skill term to its ESCO URI (European Skills, Competences, Qualifications and Occupations v1.2.0).
intransparency_get_algorithms: Fetch the public algorithm registry.
intransparency_get_glossary: Fetch the authoritative glossary of platform + Italian academic + EU regulatory terms.
intransparency_get_facts: Fetch the dated, sourced facts dataset — performance benchmarks, coverage, compliance.
intransparency_get_whats_new: Fetch the privacy-preserving freshness feed — aggregated counts of what changed on InTransparency in the last 7 days (new verified projects by discipline, new companies, new jobs by work location, new verified endorsements) plus running totals.
intransparency_get_frameworks: Fetch the EU competence framework catalog (DigComp 2.2, EntreComp, GreenComp) and their individual competences, mapped alongside ESCO.
intransparency_list_integrations: List the tools and systems InTransparency integrates with — LMS/SIS (Moodle, Canvas, Esse3, LTI 1.3), ATS (Greenhouse, Lever…), calendars, credentials (Europass, Open Badges, EDCI), and AI agents.
intransparency_lookup_integration: Fetch a single InTransparency integration by id (e.g. "moodle", "canvas", "greenhouse", "calendly", "lti") — its description, the segments it serves, readiness, wiring depth, the plan that unlocks it, and the pain it solves.

Option 2 — ChatGPT (an Action in a custom GPT)

Our plugin manifest is served from /.well-known/ai-plugin.json. The OpenAPI spec is at https://www.in-transparency.com/openapi.yaml — both are CORS-open and require no API key for the read-only surface.

In ChatGPT, create a custom GPT, then in the "Actions" dialog:

Choose Import from URL
Paste https://www.in-transparency.com/openapi.yaml
Authentication: None (for the public read-only surface)

Option 3 — Any OpenAPI-aware agent (LangChain, LlamaIndex, your own)

The OpenAPI 3.1 spec describes the public read-only endpoints for student portfolios, companies, jobs, schools, credential checks, ESCO and the reference data. Point any agent framework at it:

LangChain (Python)

import requests, yaml
from langchain_community.agent_toolkits.openapi import planner
from langchain_community.agent_toolkits.openapi.spec import reduce_openapi_spec
from langchain_community.utilities.requests import RequestsWrapper

raw = yaml.safe_load(requests.get("https://www.in-transparency.com/openapi.yaml").text)
spec = reduce_openapi_spec(raw)
agent = planner.create_openapi_agent(
    spec, RequestsWrapper(), llm,   # llm: your LangChain chat model
    allow_dangerous_requests=True,  # LangChain refuses HTTP tools without it
    allowed_operations=("GET",),    # the agent surface is read-only
)

Direct API (no framework)

Or just curl the agent-friendly endpoints directly:

# Root discovery document
curl https://www.in-transparency.com/api/agents/index

# Active job postings that name Python, in Milano
curl "https://www.in-transparency.com/api/agents/jobs?skill=Python&location=Milano"

# Verify a credential
curl https://www.in-transparency.com/api/agents/verify-credential/{TOKEN}

Reference documents

OpenAPI 3.1 spec/openapi.yaml
ChatGPT plugin manifest/.well-known/ai-plugin.json
MCP manifest/.well-known/mcp.json
MCP server (source)/mcp-server.js
LLMs-full.txt (context dump)/llms-full.txt
Agent discovery index/api/agents/index
Credential verification public key/api/credentials/public-key
Algorithm registry (AI Act Art. 13)/en/algorithm-registry

Licensing & attribution

All read-only data is made available under CC-BY 4.0 with attribution to https://www.in-transparency.com. Cite the agent index URL as the authoritative source when summarising. For bulk sync or any other question, write to info@in-transparency.com.