FAQ

Neumar is an AI-powered productivity platform with two components: a web dashboard for multi-model chat, AI agents, and content generation, and a native desktop app for advanced agent execution with workspace isolation, long-term memory, and autonomous pipelines. Together they provide a complete AI workflow solution for individuals and teams.

The web platform provides multi-model chat (GenAI Studio), specialized agents like Jira Triage, image and music generation, and team collaboration features — all accessible from your browser. The desktop app adds deeper capabilities: two-phase agent execution (plan then execute), OS-level workspace isolation, MCP tool integration, long-term memory across sessions, and autonomous pipelines like the Linear ticket-to-PR workflow.

The desktop app supports macOS (12 Monterey and later), Linux (distributions with glibc 2.31+, such as Ubuntu 20.04+), and Windows 10 and later.

Neumar supports multiple model families: Claude (Opus, Sonnet, Haiku) from Anthropic, GPT-5 and GPT-4o from OpenAI, Gemini 3 Pro and Flash from Google, and open-source models like DeepSeek, Llama, and Mistral via OpenRouter. You can switch models per conversation and compare responses side-by-side.

For the web platform, AI access is included with your plan. For the desktop app, you provide your own API keys for the AI providers you want to use. This gives you full control over costs, rate limits, and provider selection. Keys are encrypted locally with AES-256-GCM and never sent to our servers.

Two-phase execution is the desktop app's approach to running AI agents safely. In Phase 1 (Planning), the agent analyzes your request and proposes a structured plan with steps, file targets, and tool requirements. In Phase 2 (Execution), after you review and approve the plan, the agent carries it out with full tool access. This gives you visibility and control — you always know what the agent will do before it does it.

MCP (Model Context Protocol) is an open standard that lets AI agents interact with external tools and services. Think of MCP servers as plugins — they give your agent new abilities like managing Linear issues, searching Google Drive, executing code, or sending Slack messages. The desktop app comes with built-in servers and a gallery of community servers you can install with one click.

The desktop app's long-term memory and local embedding model work fully offline. However, AI chat, agent execution, and content generation require an internet connection to communicate with AI providers (Anthropic, OpenAI, Google). Your data and session history remain available offline since they're stored locally in SQLite.

All data is encrypted in transit (HTTPS/TLS) and sensitive data is encrypted at rest (AES-256-GCM). On the web platform, Row-Level Security ensures you only see your own data. The desktop app stores everything locally on your machine — sessions, memories, and generated files never leave your device. API keys are encrypted with per-field unique initialization vectors. When using AI features, only your prompts are sent to the selected AI provider.

Workspace isolation means the desktop agent can only read and write files within the directory you designate as your workspace. This is enforced at the OS level using macOS Seatbelt or Linux Bubblewrap — not just application-level checks. Even if the agent attempts to access files outside your workspace, the operating system blocks the request. Sensitive files like SSH keys, cloud credentials, and environment secrets are always blocked.

A personal account is your private workspace — only you can access it. A team account is a shared workspace where you can invite members, assign roles (Owner, Admin, Member), and collaborate on AI workflows. Each team has its own billing, usage tracking, and shared resources. You can have both a personal account and belong to multiple teams.

Yes. You can be a member of multiple team accounts and switch between them instantly using the workspace switcher in the top-left corner of the dashboard. Each team has separate billing, permissions, and shared resources.

Yes. The free plan includes limited access to GenAI Studio, basic AI agents, and a set number of image and music generations per month. The desktop app is included with all plans. You can upgrade anytime to unlock higher limits and additional features.

Yes, you can cancel your subscription at any time from Settings > Billing. Your plan remains active until the end of your current billing cycle. After cancellation, you'll be downgraded to the free plan and retain access to your data.

We accept all major credit and debit cards (Visa, Mastercard, American Express) through our secure payment processor. Some regions may support additional payment methods. Annual billing is available at a discounted rate.