Common questions.

A curated marketplace of 100 hand-authored AI agents that install into Claude Code, ChatGPT, OpenAI Assistants, Cursor, n8n, Zapier, or anything that takes a system prompt. Pick one. Click Install. Drop into your tool. Ship work.

Most agent catalogues are scraped or LLM-generated. Empty system prompts. No point of view. Identical tone. We wrote each agent's prompt by hand, with concrete decision trees, refusal rules, and output formats. They have opinions because they were made by people.

For browsing and exporting, no. For running an agent, yes — you bring your own LLM credentials. AgentDash gives you the system prompt and the tool config. The model call happens on your side, so your data never touches us.

Each agent has a recommended model based on the work it does (Opus for heavy reasoning, Sonnet for balanced, Haiku for fast, GPT-4 variants where they fit). You can swap any of them for a different model. The system prompt is portable.

One agent: Personal Chief of Staff. It's the only free agent and it stays free forever. Browse all 100 agents without an account, but install only the free one.

$19/month. All 100 agents minus the 16 Enterprise ones — 84 agents total. All 8 export formats, the workflow builder up to 3 chained steps, favorites, and install history.

$79/month. Every single agent including the high-stakes ones (Smart Contract Auditor, M&A Analyst, Tax Strategist, Compliance Auditor, Threat Modeler, etc.), unlimited workflow steps, programmatic install via API, 5 team seats, white-label export, and the right to submit custom agents to be added to the catalog.

Yes. Cancel inside Settings. Access continues to the end of the billing period. No refunds for partial months.

Yearly plans get the equivalent of two free months. Educators and registered nonprofits get 50% off Pro — email hello@agentdashapp.com with proof.

Eight formats: Claude Code subagent (.md with YAML frontmatter), OpenAI Assistant JSON, ChatGPT Custom GPT, raw system prompt, Cursor .cursorrules, n8n workflow node, Zapier prompt block, and a full Markdown bundle.

We track the format spec for each tool and update the export template when they change. If your tool deprecates a format, the export adapts and existing files keep working until you re-export.

Of course. The export is yours to modify. We recommend keeping a copy of the original alongside your edits so you can diff against future updates from us.

No. They produce informational guidance to help with research, structure, and drafting. They explicitly disclaim being a substitute for licensed counsel, a registered tax preparer, or a clinician. For binding decisions in those domains, talk to a qualified professional.

We don't see it. The model call happens on your side with your credentials. AgentDash only stores your account information, subscription state, and which agents you've installed.

Independent operators who got tired of agent catalogues full of empty templates. We wrote 100 prompts the way we wished someone else had. More on the About page.

Yes — email hello@agentdashapp.com. Enterprise customers get priority on suggestions.