All Features

Every HivemindOS guide, grouped by what you want to do. New to HivemindOS? Read Getting Started first.

Talk to agents and get work done

• Agents and chat: pick an agent, attach files or a folder, and continue earlier conversations.
• Work and automations: track tasks on a shared board, review results, and schedule work that repeats.
• Apps using HivemindOS: let Codex, Claude Code and other apps on your computer use its browser, brain and computer control, with your permission.
• Capability approvals: review the tools an agent proposes before a build starts.
• Agent-native workflows: see what an agent will do, is doing, and did.
• Slash commands: shortcuts you can type in any chat.
• Agent plugins: load portable skills and MCP servers from plugin packages.
• Computer interaction: let an agent operate apps on your computer, with your approval.

Your machines and setup

• My Hive: see your computers and agents, add a machine, and fix setup problems.
• Machine permissions: control what agents may do on each computer.
• HivemindOS Machines: ready-to-use cloud virtual machines for your hive.
• Settings, files, and alerts: shared setup, approved files, alerts, and health checks.
• Desktop app and browser extension: the ways to use HivemindOS.

Shared knowledge and skills

• Hive Superbrain: shared memory, knowledge, skills, and setup for every permitted agent, without putting secrets in the vault.
• Documents and Brain Drop: give agents a document for one chat, or add it to the brain for later.
• Hivemind Sync: keep the brain, shared setup, and handoffs available on all your trusted machines.
• Handoff transfers: leave files for one specific agent or machine.
• Skills: reusable procedures agents can follow, plus optional skill packs.
• Hive Fusion: turn a prompt into a reusable skill built from what your hive already has.
• Memory benchmarks: measured recall quality, speed, and token savings.

Build and create

• App Builder and hosting: describe an app, preview it, and publish it.
• Zero Human Companies: give a team of agents a business goal, a budget, and approval rules.
• Founder Mode: turn one outcome into a reviewable company blueprint.
• Web research: search, read, extract, and cite public web pages and PDFs.
• Connected intelligence: follow source changes and keep sensitive work on your devices.
• SRE investigations: package an incident as redacted evidence for root-cause analysis.
• Photo library: index large photo folders with reviewed AI keywords.
• Analytics: privacy-first web analytics for sites you own.

Communicate and assist

• Calling: talk with agents by voice.
• Socials: draft, approve, and publish posts on connected accounts.
• Inbox triage: turn rough captures into clean notes and tasks.
• Bookings: let an agent manage your calendar appointments.
• X command bot: mention the HivemindOS bot on X to save, analyze, or reply.
• Office companion: create and edit Word, Excel, and PowerPoint files with agents.
• Page Agent: an experimental agent that works inside a web page.
• Read Along: read a document and ask an agent about any passage.
• Privacy Suite: privacy-focused apps your agents can use.
• Beeline family profiles: separate identities for each person you help.

Money and compute

• Wallets and spending: balances, budgets, approvals, and paying for services.
• Finances: accounts, budgets, and goals in one private workspace.
• HONEY for X posts: how posting about HivemindOS earns recognition.
• Trading: crypto, stocks, perpetuals, and prediction markets, with safety limits.
• Agent-analyzed copy trading: compare a copy-trader with a research-assisted twin.
• Hivemind Bots: always-on cloud agents that keep working while your computers are off.
• Hive Compute: earn from models you host, or use community-hosted models.
• Token and cost savings: how HivemindOS cuts token use and model spend.

Connections

• Integrations: every app and service you can connect.
• Apps and agent providers: connect outside tools and specialist agent services.
• Integrations and work history: connected services and the record of finished work.
• MiroShark and runtime gateways: rehearse a scenario before running it for real.

Labs

Experimental and research features. Trading labs use paper trading unless the page says otherwise.

• Agent evaluations: how agent work is checked against evidence.
• Agent challenges: one bounded goal, one shared board, one evidence trail.
• Loop engineering: repeatable agent loops with gates, budgets, and receipts.
• Evo optimization: automated experiments that try to improve a measurable result.
• Harness experiments: compare agent setups without mistaking cheaper for better.
• Quant research swarm: test market ideas and keep the evidence.
• Signal Trader: trade one measured pattern, with practice and real money kept separate.
• KOL Radar: see which crypto accounts on X move your token.
• Penny-stock paper lab: study one limit-order method without a broker.
• Prediction-market learning lab: a research-only experiment that never places a trade.