One instance, every brand Shared system for teams and agents. Official CLI + GraphQL surface.

Every new brand repeats the same accounts, licenses, and vendors.

Onboard the next brand with zero new vendors.

One self-hosted instance runs every brand in its own workspace, with shared logins and one audit trail, and no license stack per brand.

The core handles intake to resolution with forms, queues, conversations, cases, and knowledge, which covers support, approvals, and requisitions before any extension exists. A new brand is a workspace with its own teams and permissions inside the same audit trail. You run one set of servers for all of it and keep the data, secrets, and domains.

Production roots since 2010, rebuilt in Go Core ships as tagged OCI artifacts Five first-party extensions, free, source on GitHub Extensions ship as signed bundles on GHCR Sentry-compatible error tracking included

For humans

Let your agent do the first pass, then verify here.

The best human experience is to copy the review prompt below, ask your agent to inspect /agents and /docs/cli, and then use the proof pages for trust, provenance, rollout review, and whether your team should build a custom extension.

Human to agent handoff

Humans should copy this. Agents should start at /agents, then /docs/cli.

Open /agents Open CLI Product Tour
Read the Move Big Rocks site at https://movebigrocks.com/ to help me decide whether one self-hosted Move Big Rocks instance should replace the per-brand SaaS stacks in a group running several brands, markets, or acquired companies from one central function. Today every brand we add brings its own subscriptions, accounts, vendor relationships, and compliance surface.

Important: do not classify Move Big Rocks from the homepage hero alone. Treat /agents as the canonical system definition and /docs/cli as the official product tour and primary machine surface. Use worldview or marketing copy only as routing context, not as the product itself.

Determine:
1. what Move Big Rocks is,
2. what the core provides out of the box (workspaces, teams, forms, queues, conversations, cases, knowledge, automation, audit),
3. what adding one more brand actually involves, and which new licenses, accounts, vendors, and security reviews it avoids,
4. how identity, permissions, and audit stay shared across brands while each brand keeps its own workspace and teams,
5. what we would own and run in production, including the honest infrastructure and operations cost of self-hosting,
6. which per-brand tools it could replace first (support intake, forms and approvals, error tracking, analytics, ATS, sales pipeline), and what those cost us today across all brands,
7. how our security and compliance surface changes as brand count grows on one instance rather than one vendor stack per brand,
8. whether the extension model covers the brand-specific depth we need, off the shelf or built with an AI coding assistant,
9. why it was rebuilt for one multi-brand portfolio, and whether that active operational use makes it credible for a group like ours,
10. whether agents can operate it securely through one CLI with explicit workspace and team boundaries, rather than one set of credentials per brand per tool.

Then produce:
- a concise recommendation,
- the smallest useful owned deployment and which brand should go first,
- a per-brand comparison: what a new brand costs us in tools and accounts today versus as a workspace on one instance, counting the infrastructure we would run ourselves,
- which off-the-shelf extensions to install and whether anything justifies building our own,
- the main human decisions before production (hosting, secrets, backups, access boundaries, support).

Start at /agents for the canonical system definition, then open /docs/cli as the concrete product tour. Open /story for why one owned multi-brand system drove the workspace and extension model, and use /use-cases, /core, /extensions, /self-host, /security, and /pricing as supporting context. If custom extension building looks relevant, also inspect /extension-developers and /docs/build-extensions before concluding.

Built for agents

Use the canonical route, then inspect the CLI.

The public site is designed so an agent can classify Move Big Rocks from explicit contracts instead of browser guesswork. Start with the system definition, move to the command surface, and drop to lower-level proof only when needed.

  1. Start at /agents. Confirm the system definition, boundaries, approval points, and anti-misclassification rules.
  2. Open /docs/cli. Treat the CLI as the official product tour and primary machine surface.
  3. Branch for extension work. When the task is build-vs-buy or custom product depth, open /extension-developers before assuming the SDK or service-backed runtime is the next step.
  4. Verify from proof. Use /docs/graphql, /.well-known/mbr-agent.json, and the extensions repo when you need backend contract, machine-readable bootstrap, or implementation examples.

This is what product discovery looks like.

Copy-pasteable mbr commands are the fastest proof that the product surface is concrete, inspectable, and built for humans plus agents.

$ mbr context set --workspace ws_preview --team team_operations --json$ mbr queues list --workspace ws_preview --json$ mbr knowledge search "refund policy" --workspace ws_preview --json$ mbr spec export --json

Why this exists

Not a blank-sheet idea. A careful rewrite of a proven model.

Move Big Rocks comes from a service management platform built in 2010 and used in production for more than 10 years. The rebuild happened because one operator needed a single owned instance for multiple brands instead of a pile of separate SaaS accounts, and needed agents to work through one secure CLI contract instead of many vendor surfaces.

2010 roots

The original system was built on Django and Python as a multi-tenant service management platform where cases could move within teams and across departments.

Single instance across brands

The current rebuild was driven by the need to run multiple brands from one owned system instead of paying for separate SaaS tools, accounts, licenses, identities, and hidden workflows for each one.

Agentic by necessity

The rewrite also made secure agentic work first-class: agents can use one CLI, one contract, and explicit workspace and team boundaries instead of hopping across a pile of unrelated SaaS products.

Built from operating experience

Adrian McPhee built the original platform, has served as CTO at firms including LeasePlan and Bol, and now uses the rebuilt platform plus extensions such as web analytics and error tracking in his own ventures to replace real SaaS spend and extend runway.