The questions behind company-wide AI.

HQ works with them. It is not another foundation model and does not require your team to abandon the AI tools it prefers.

Claude Code, Codex, and Cursor can work directly with HQ Core. HQ MCP aggregates the company’s approved memory, tools, and integrations into supported chat products. HQ Agents can use the same context and connections while working persistently in channels such as Slack and email.

HQ organizes the information AI needs into ordinary, inspectable company files: knowledge, Skills, Projects, Workers, and Policies. Teams can add source documents, capture approved learnings, connect systems, and turn repeated work into reusable workflows.

HQ does not need to treat every conversation as permanent truth. Your team decides what should be captured, who can see it, and when it should be updated or retired.

A Skill is a reusable workflow for a job your team does more than once. It can include instructions, source material, tools, quality checks, and rules for handling the result. A teammate can invoke that workflow instead of rebuilding the prompt and process from memory each time.

Because Skills are shared company files, they can be inspected, improved, tested, and distributed to the people or Agents that need them.

A Worker is a job definition: a reusable role with instructions, Skills, context, and a clear definition of done. Anyone on the team, or any Agent, can run that Worker for a specific task. It does not have its own login, inbox, or monthly bill.

An HQ Agent is a persistent teammate. It has an identity, company membership, scoped access to files and Secrets, and its own runtime, so it can work in Slack, email, or on a schedule. Agents use Workers the way people do: the Worker is what to do; the Agent is who does it.

Nontechnical teammates can use HQ without becoming command-line experts. The most common path is Claude Desktop or the ChatGPT app with HQ MCP: company memory, Skills, and connected systems show up in a familiar chat, and that is usually a great experience. HQ Agents can also work alongside the team in Slack and email, and HQ Deploy can publish apps, reports, and dashboards that people simply open and use.

A smaller operator group will usually install HQ, build Skills, manage Integrations, and govern access, while the broader team uses approved workflows through the interface that fits its work.

Start by installing HQ, creating or joining your company, enabling Sync, and opening the HQ workspace with a supported AI coding tool. Guided setup helps you add company context, configure permissions, store Secrets, connect Integrations and MCP servers, and create the first shared workflow.

From there, activate the products you need, such as hosted MCP, Agents, meeting capture, and Deploy. Team plans include a white-glove launch to assess workflows, configure HQ, train the team, and drive adoption.

Startup is free for up to 50 people with no seat fees. Team is $500 per month as a flat company subscription and includes three AI Agents, hosted MCP, meeting and company-intelligence features, custom Deploy domains, and white-glove launch. Enterprise starts at $10,000 per month for requirements such as SSO, dedicated infrastructure, and enterprise support.

Optional usage is priced separately where applicable: additional Agents are $100 per month, Outposts are $100 per month, and meeting capture is $1 per hour. Check the pricing page for the latest plan details.

HQ Secrets + Integrations is the shared connection layer for company systems. Connect a supported service or custom MCP server once, keep the credential in the encrypted vault, and provision that connection to the people and Agents allowed to use it. HQ MCP can aggregate those approved connections with company context and make them available through supported AI clients.

Integration depth varies by system. Some connections provide ready-to-use search and actions, some expose a service through MCP, and some require a custom implementation.

You do not need to be a Git expert to use an HQ-powered workflow. Nontechnical teammates can use approved Agents, chat connections, and published outputs without managing branches or pull requests.

Git is a technical dependency of the full local HQ workspace and is useful for history, review, and collaboration. HQ does not automatically publish your company files to GitHub. Your team controls what is shared and where it goes.

HQ is local-first. HQ Core keeps the working company brain in ordinary files in the HQ workspace, where the team can inspect and control it. When a company uses HQ Cloud, approved data needed for cloud features is synchronized into a company-scoped tenant.

Cloud data is encrypted in transit and at rest. Credentials receive separate field-level encryption, and company access is isolated at the application and query layers. The managed service currently uses AWS S3 for meeting and file storage and MongoDB Atlas for application data.

Every company has its own tenant. Within that company, access can be assigned to people, groups, and Agents according to what they need to do. Grants can distinguish among company files, Secrets, Integrations, MCP connections, Agent capabilities, and deployed outputs.

The recommended approach is least privilege: start with the smallest useful access and expand a person’s or Agent’s permissions only when the work requires it.

HQ Secrets + Integrations is the company’s API key and credential manager. Secrets belong in its encrypted company vault, not in prompts, documentation, chat transcripts, repositories, or shared environment files.

Access can be granted to the specific people, groups, and Agents that need a credential. HQ can inject the approved Secret at runtime so a tool can authenticate without copying the raw value into the company brain.

HQ Cloud keeps the approved parts of HQ available to the people and Agents that have access. That includes company memory, Skills, Projects, Workers, Policies, MCP connections, Secrets, and Integrations.

When changes cannot be merged safely, the conflict is surfaced for review rather than silently erasing one version. Permissions still determine who can read or write each shared resource.

Skills are ordinary files, so their instructions and supporting material can be inspected, compared, tested, reviewed, and rolled back. A team can name an owner, test a change against representative work, review the result, and sync the approved version.

Teams can use lightweight review for low-risk Skills and stricter human approval for workflows that publish, spend money, change systems, or contact customers.

Each company is a separate trust boundary with its own knowledge, members, Policies, Integrations, and credentials. Access to one company is not permission to read or act inside another company.

Agencies and multi-company operators should keep each client in the correct company workspace and grant access deliberately. Inside a company, narrower groups and resource permissions can limit access further.

HQ currently provides guided desktop installation for macOS and Windows. The underlying HQ shell and file-based workspace also support Linux, macOS, and Windows through Git Bash, although installation and feature parity may differ by platform.

Server and headless environments should be evaluated separately from the desktop onboarding path.

HQ is a hybrid, local-first product suite. HQ Core and the working company files live in a local workspace. HQ Cloud manages synchronization, sharing, membership, and access. Secrets + Integrations, HQ MCP, HQ Agents, HQ Deploy, meeting capture, and the intelligence layer use managed services when enabled.

Enterprise is the path for companies with dedicated infrastructure, data-residency, private-network, or specialized deployment requirements.

HQ Core is the open-source, file-based foundation: the structure for company knowledge, Skills, Projects, Workers, and Policies that AI tools can use locally.

The managed suite adds HQ Cloud for sync and governed access, Secrets + Integrations for credentials and company-system connections, HQ MCP for bringing those capabilities into chat, HQ Agents for persistent AI teammates, and HQ Deploy for publishing the work AI creates.

Yes. HQ Agents is the managed product for deploying persistent AI teammates with a role, memory, tools, responsibilities, and team channels. A company is not limited to those packaged Agents.

Existing or custom Agents can use HQ through the local file structure, HQ MCP, provisioned Integrations, or a custom implementation. Each implementation should define the Agent’s identity, data scope, tool permissions, Secret access, and human approval points.

When HQ meeting capture is enabled, the meeting bot can create a transcript and structured meeting record inside the company brain. Existing meeting notes and transcripts can also be imported through supported paths.

Authorized teammates and Agents can search the conversation, recover decisions, prepare follow-ups, create briefs, identify recurring objections, and feed approved learnings into Projects or reusable workflows.

You do not have to start with a blank company brain. Existing files and documentation can be added to company knowledge, and supported systems can be connected or imported. Repeated prompts, checklists, and operating procedures can become shared Skills.

The best rollout is curated rather than indiscriminate: start with authoritative sources, remove Secrets and obsolete material, preserve ownership, and assign the right permissions.

HQ Deploy can turn a finished report, dashboard, portal, site, or lightweight application into a hosted URL without making every recipient install the HQ workspace.

The creator chooses the appropriate access mode: public, password-protected, or authenticated for the company, depending on the output and plan. Team plans currently include custom Deploy domains.

An administrator invites a person into the correct company and assigns the role, groups, files, tools, and Secrets needed for that person’s work. Membership does not have to mean unrestricted access to every resource.

Clients should receive narrowly scoped access inside the appropriate company or work in a separate client company when stronger isolation is required.

HQ Agents can work alongside the team through Slack and supported email channels, using the same company memory, Skills, tools, and permissions as the rest of HQ. HQ MCP also brings approved company context and connected systems into supported chat clients.

Mobile access generally comes through the mobile version of a supported channel such as Slack, email, or chat rather than a separate full-featured HQ mobile application.

HQ governance starts with visible company files, tenant isolation, scoped permissions for people and Agents, protected Secrets, company Policies, and activity history.

Human review and approval gates can be built into sensitive workflows, especially before publishing, spending money, changing a system of record, or contacting a customer. Enterprise requirements can be evaluated during implementation.

HQ is a software product suite: open-source HQ Core plus the managed HQ Cloud, Secrets + Integrations, HQ MCP, HQ Agents, and HQ Deploy products. Meeting capture, ontology, signals, search, and activity capabilities extend that shared system.

Paid plans can include hands-on launch, configuration, training, and support. Custom implementation may be available for unusual integrations or deployment work. Strategic advisory is a separate Indigo service.

HQ gives AI a more reliable starting point: current company files, approved sources, explicit Policies, reusable workflows, scoped tools, and permission-aware context. A workflow can require the model to search the right sources, show receipts, distinguish facts from inferences, and ask for review before a sensitive action.

HQ reduces avoidable guesswork, but it cannot eliminate model errors. Critical claims and actions should remain traceable to source material and use human review where the cost of being wrong is high.