HomeAbout UsPricingContact Us
FrameworksISO, SOC 2, NIST & more, explainedBlogArticles from the security deskDocumentationProduct guides & how-tosAPIBuild on the Platform APIMCP integrationConnect your AI to your workspaceFrequent questionsAnswers, straight
Log inBook a demo
HomeAbout UsPricingContact Us
Resources
FrameworksBlogDocumentationAPIMCP integrationFrequent questions
Log inBook a demo
Ready when you are

Be audit-ready by default.

Start free with six frameworks, thirty policies, and one living picture of your security program.

Get started freeTalk to an expert

Cyber Security Framework as a Service: governance, risk and compliance, run from one living platform.

Compliance insights, monthly. No spam.

Product

PlatformPricingRequest a demoAccess CSFaaS

Resources

FrameworksBlogDocumentationAPIMCP integrationFrequent questions

Company

About usContact usDarkProtect, managed services

Legal

Privacy policyTerms & conditions
CSFaaS is operated by Darkprotect (GIB) Limited, registered in Gibraltar (company number 125185). Registered office: Sovereign Place, 117 Main Street, GX11 1AA, Gibraltar.© 2026 CSFaaS, All rights reserved.All systems operational
Resources ▸ MCP integration

Connect your AI to CSFaaS.

Add one server URL to Claude, Claude Code, Claude Desktop or any MCP client, sign in through your browser, and ask GRC questions in plain language. Grant read & write and your AI can act on the answers too. No API key to paste. The MCP server holds no database credentials; every tool call runs as you, under the same row-level security as the app.

Get started freePrefer raw HTTP? API docs

csfaas mcp · connection

endpoint   https://mcp.csfaas.com/mcp
transport  Streamable HTTP
auth       OAuth sign-in (browser)
access     read by default · write by grant · runs as you, under RLS
What you can ask

Plain questions, grounded answers

Your AI calls tools against your live workspace and answers from real data, not from a stale export. Ask what you would ask a colleague, and with a read & write grant, delegate the follow-up too.

your ai client

connected to csfaas
What are my top 5 risks?
search_riskssort: score · limit: 5

Your highest exposure is RSK_000008, "Unpatched edge VPN", scored 20 of 25. Four more follow, and three of them still have no remediation plan.

Global GRC queries

Ask across the whole program: risks, controls, frameworks, vendors, systems and reviews.

Risk Assessment Assistant

Pull a risk or a demand with its full assessment and let your AI critique scoring, wording and treatment.

Remediation review

Check documents and evidence against remediation plans, their owners and their due dates.

Audit evidence

Walk audit questions, read their threads and verify that evidence answers what the auditor asked.

Security policies

Read policy trees and check what a policy actually commits you to.

Setup guides

Connected in two minutes

Add the server URL to your client, approve it in your browser, done. Pick your client below.

0

Add the server, then authorize in your browser. Point your client at https://mcp.csfaas.com/mcp. On first use it opens a browser where you sign in to CSFaaS and pick the workspace, role(s) and access level (read, or read & write) it may act as. No API key to paste.

Run this once in your terminal. No header, no key. Start a session, and the first time it uses the server your browser opens to sign in and authorize it.

terminal

claude mcp add --transport http csfaas https://mcp.csfaas.com/mcp

The first time your client reaches the server it opens a browser to sign in to CSFaaS. Approve the workspace, role(s) and access level, and you are connected. No key to paste, no token to manage.

Prefer an API key for scripts or REST? Create one in Settings ▸ API and send it as a Bearer token. That path is for headless and CI use, not for connecting an AI client.

Tools reference

Read and write across every module

Everything the server exposes to your AI, grouped into suites. Expand a suite to see its tools. Reads return the same rows, and only the rows, you can see in the product; write tools require a read & write grant and act as you.

15 tools · reference
READ

whoami

Who the key belongs to, their workspace and role

READ

search_risks

Search the risk register by text, score or status

READ

search_controls

Search the control library, including evidence gaps

READ

list_frameworks

Deployed frameworks with their versions

READ

framework_statistics

Implementation statistics for a framework

READ

list_third_parties

Vendors with tiering and risk posture

READ

list_systems

Systems and assets with their classification

READ

list_reviews

Periodicity reviews, owners and due dates

READ

search_catalog_controls

Search the control and threat catalogs

READ

get_risk

One risk with its full assessment and treatment

READ

get_demand

One risk demand with status and assessments

READ

read_evidence

Read an evidence record or linked document

READ

list_audit_questions

Questions of an audit campaign

READ

get_audit_question_thread

One audit question with its full thread

READ

list_policies

Policies with their current version and state

Empty means out of scope

An empty result means nothing in your authorized scope. Row-level security filters invisible records out; it does not raise an error, and your AI should not treat it as one.

Trust & FAQ

Safe by construction

The server is a thin, credential-less gate in front of the same authorization layer the app uses. Here is what that means in practice.

Read by default

A connection starts read-only: your AI can read, reason and propose. Grant it the read & write access level and it can also create, update and link records, always as you, never beyond your permissions.

Your permissions, exactly

Every tool call is authorized by the same PostgreSQL row-level security as your session in the app. Different role, different answers.

Instant revocation

Revoke a connected client under Settings ▸ API (Connected AI clients), or an admin flips the workspace API kill-switch, and access dies immediately. No cache, no grace window.

Any MCP client

Streamable HTTP with browser OAuth. Claude, Claude Code, Claude Desktop, or any client that speaks MCP.

401

Not authorized yet

Your client is not signed in yet. It opens a browser to sign in to CSFaaS and authorize it; approve the workspace and role(s), and it reconnects on its own.

403

No access

The signed-in user lacks permission for this resource, the connection is read-only and the tool writes, or a workspace admin disabled API access.

429

Rate limited

Too many tool calls in a short window. Your client should back off and retry; steady conversations never hit it.

Only if you let it. A connection is read-only unless you grant it the read & write access level; without that grant, any write is refused. When you do grant it, your AI can create, update and delete records, run demand workflow actions, link records, comment and attach evidence. Every change runs as you, under the same row-level security as the app, and every change is logged. A workspace admin can also disable writes, or all API access, for the whole workspace at any time.

No. It holds no database credentials and keeps no copy of your data. Your browser sign-in (or an API key, for scripts) resolves to a short-lived, workspace-pinned token, and the database authorizes every request, reads and writes alike.

An empty result means nothing in your authorized scope. Row-level security filters invisible records out; it does not raise an error.

Any MCP client that supports Streamable HTTP with OAuth: Claude, Claude Code, Claude Desktop, and most agent frameworks. The generic mcp.json block covers the rest.

Ask your program anything

Your GRC data, one question away

Add the server URL to your client, approve it in your browser, and ask your first question in under two minutes. Included in every plan, reads and writes alike, at no extra cost.

Get started freeRead the API reference