Early access · now onboarding teams
See everything you run.
Know what breaks before it does.
Merethold is a living, auto-discovered map of your servers, containers, networks and services, connected to the products and customers that depend on them. Ask “what breaks if this goes down?” and get an answer in seconds.
- Read-only integrations
- MCP server for AI assistants
- Built in the UK
-
20+
read-only integrations
monitoring, cloud, DNS, containers, billing and more
-
500+
assets on one map
with 1,000+ relationships, tested for smooth navigation
-
15 min
default SSH certificate life
short-lived access for people and AI agents
-
0
write access required
integrations only read, by design
The problem
Your infrastructure lives in twelve tabs. Nobody sees the whole picture.
Metrics in one place, servers in another, the asset list in a spreadsheet nobody trusts, and the knowledge of which customer depends on which database in someone’s head. When something fails at 2 a.m., that gap is the outage.
Merethold pulls it into one living model, discovered from the tools you already run and kept current automatically, so the answer to “what does this affect?” is a click, not a meeting.
The living map
From the whole estate to a single container in one scroll.
- 01
The whole estate, live
Every server, network, DNS zone, load balancer and certificate, laid out by region and network, coloured by live health.
- 02
Open any server in place
Zoom straight into app-prod-01: specs, addresses, live CPU, memory and disk from Grafana, without losing your place on the map.
- 03
See what runs inside
Containers, databases, certificates and backups discovered from Docker, Caddy, Prometheus and your object storage.
- 04
Know who depends on it
Products, customers, owners and the right runbook, one hop away. That is the context you need at 2 a.m.
Impact analysis
What breaks if this goes down? Ask before it happens.
Merethold walks the dependency graph from any component to the applications, products and customers that rely on it, and it knows the difference between a failover you have proven and one you are hoping for.
Take a component down
gw-edge-01 → gw-edge-02 failover was confirmed by change CHG-0012.
- Applications
- —
- Products
- —
- Customers at risk
- 0
Integrations
Discovered from the tools you already run.
Connect a read-only token and Merethold builds the map for you, then keeps it current. Manual notes and layouts survive every re-sync, and anything that disappears is flagged for review, never silently deleted.
Supported integrations: Grafana, Prometheus, OVHcloud, S3 object storage, Cloudflare, Cloudflare Zero Trust, WireGuard, Docker, Caddy, GitHub, Sentry, Uptime Kuma, Better Stack, UptimeRobot, Grafana OnCall, PagerDuty, GitHub Issues, Linear, Jira, Stripe, Xero, QuickBooks, Email providers.
One workspace
More than a map. The operating record of your business.
-
Live health and telemetry
Up/down, CPU, memory, disk and firing alerts on every asset, each value stamped with its source and age, so stale data is never mistaken for healthy.
-
Measured link traffic
Per-link throughput from real interface counters, drawn on the map.
-
Runbooks and procedures
The right procedure opens with the asset in context. Runs are recorded step by step.
-
Incidents and changes
Snapshot impact when an incident opens; record drills and changes as failover evidence.
-
Renewals and costs
Domains, certificates, servers and subscriptions, with expiry dates and what each one pays for.
-
CRM, connected
Customers ↔ products ↔ infrastructure. Know which accounts to call before they call you.
-
Backups you can trust
Snapshots, the last restore test and the restore procedure, right next to the database they protect.
AI-ready operations
Let your AI assistant read the map. Keep a human on the keys.
Merethold’s built-in MCP server lets Claude and other assistants query assets, topology, health, impact, incidents and runbooks with a personal, revocable key. When an agent needs a shell, it asks for a short-lived SSH certificate and your rules decide what happens next.
What breaks if db-prod-01 goes down tonight?
● merethold.impact_analysis(asset: "db-prod-01")
⎿ 3 applications · 2 products · 41 customers at risk
db-prod-01 is the primary database behind app-prod-01, -02 and -03. A replica runs on db-prod-02, but failover has never been confirmed, so treat all three as at risk.
Affected: Harbour Pay (28 customers) and Fieldbook (13). Runbook: RB-007 Promote replica.
Restart billing-api on app-prod-02 for INC-0142.
● merethold.request_ssh_access(server: "app-prod-02", tier: "fix")
⎿ Pending approval · request SSH-0311
SSH-0311 · fix tier
merethold-fix@app-prod-02
- Request receivedPublic key only; private keys are refused.
- Policy checkFix tier: open incident assigned to the requester. Passed.
- Approved by an ownerApproved in Merethold with one click, and audited.
- Certificate issuedOne server, one login, source-address bound.
MCP server built in
Streamable HTTP endpoint with personal keys that are hashed at rest, expire on your schedule and revoke instantly.
Look, Fix, Admin tiers
Low-risk read access can be automatic. Fix needs a matching incident. Admin always waits for a human. Certificates last 15 minutes by default.
Every login traceable
Certificate key IDs tie each SSH session back to the request, reason and approver, with daily rate limits that pause runaway agents.
Security and trust
Built so your credentials never have to leave home.
Merethold sees your infrastructure without being able to change it. The Merethold Agent goes further: it runs inside your network, talks outbound only, and keeps internal credentials where they belong.
your network · 10.20.0.0/16
- grafana.internal.example
- docker hosts × 6
- caddy · edge proxy
- credentials stay here
merethold cloud
Receives inventory and health summaries, never your internal tokens.
-
Read-only by design
Every integration in this release only reads. Merethold never needs write access to your cloud, DNS or containers to build the map.
-
Encrypted credentials
Tokens are sealed with AES-256-GCM, bound to your workspace, or kept as references to your own vault. They are never returned, logged, exported or audited in clear.
-
Tenant isolation and roles
Every query is scoped to your workspace. Owner, admin, engineer, CRM and viewer roles, and CRM staff see impact without IPs or network detail.
-
Full audit trail
Changes, imports, approvals and access are recorded with who, what and when. Layout tweaks are logged separately from real configuration changes.
How it works
Connect. Discover. Understand.
No agents to install on every server, no diagrams to draw by hand. Point Merethold at what you already have.
-
01
Connect
Add read-only tokens for Grafana, your cloud, DNS and more, or run the Merethold Agent inside your network. Import a CSV or JSON inventory if you have one: dry-run first, no duplicates on re-import.
-
02
Discover
Merethold reconciles every source into one model of assets and relationships, grouped by region, network and subnet. Each fact keeps its source and timestamp.
-
03
Understand
Trace dependencies, run impact analysis, attach runbooks, owners and customers, and confirm failover with evidence. Your team, CRM staff and AI assistants all work from the same picture.
What exactly is Merethold?
Merethold is a topology-first infrastructure and business-operations platform. It builds a living, editable map of your servers, containers, databases, networks, DNS, load balancers, certificates and backups from read-only integrations, then connects that map to your products, customers, runbooks, incidents, changes, renewals and costs.
Does Merethold need write access to my infrastructure?
No. Every integration in the current release is read-only. Merethold reads inventory, health and configuration to build the map; it does not change your cloud, DNS, containers or proxies. The only capability that grants access to servers is the optional SSH certificate authority, which issues short-lived certificates under rules and approvals you control.
How does impact analysis decide what is affected?
Merethold walks the dependency graph from the component you select to the applications, products and customers that rely on it. Redundant paths are taken into account, but only failover you have confirmed with evidence (for example a drill or change record) counts as protection. Unconfirmed secondaries are flagged as risk, not assumed to work.
Can I use it with Claude or other AI assistants?
Yes. Merethold includes an MCP (Model Context Protocol) server. Create a personal key in settings and your assistant can query assets, topology, health, impact, incidents, changes, renewals, backups and runbooks. Keys are stored only as hashes, can expire and can be revoked instantly.
What is the Merethold Agent?
A small service you run inside your own network. It reads internal systems such as Grafana, Docker hosts and proxies, and sends inventory and health summaries to Merethold over an outbound-only HTTPS connection. Internal credentials stay on your side, and no inbound ports need to be opened.
Can non-technical colleagues use it?
Yes. Roles include owner, admin, engineer, CRM and viewer. CRM and account staff can see which customers an incident affects and what a customer relies on, without seeing IP addresses, interfaces, firewall rules or access notes.
How big an estate can it handle?
The map is tested with workspaces of more than 500 assets and over 1,000 relationships, with level-of-detail rendering, grouping by region and network, and keyboard navigation throughout.
How much does it cost?
Merethold is in early access. We are onboarding teams individually and agreeing pricing with each of them. Get in touch and we will talk through your estate and what you need.
See your whole estate by the end of the day.
Connect a read-only token, watch the map assemble itself, and ask your first “what breaks if…” question before lunch.
Already have an account? Sign in