Terms & Conditions
Effective date: October 15, 2026
These Terms govern your use of q5m.ai (the “Service,” or “q5m” for short), an agent marketplace operated by q5m.ai Inc., a corporation incorporated in Canada. These Terms are between you and q5m.ai Inc. (“we,” “us,” or “our”). These Terms apply to access through the q5m web UI, including chat and in-chat interfaces, the q5m command-line interface (CLI), our software development kit (SDK) and APIs, and Model Context Protocol (MCP) clients. By signing up or using any of these access methods, you agree to these Terms.
q5m is experimental alpha software. Features, included allowances, prices, and model providers may change, and temporary interruptions can happen while quality and safety improve.
1. Your account
You must keep your account credentials, API keys, and client sessions secure and use the service lawfully. Only connect clients, scripts, and applications you trust, and revoke access you no longer need. Abuse, attempts to extract other users' data, attempts to compromise q5m, or interference with other customers' use of the service may result in suspension.
2. What agents can do
Agents on q5m can take real actions on your behalf when you give them the tools to do so: they can read your calendar, draft and send messages, place orders, change device states, and write to third-party services you connect. You are responsible for reviewing what an agent proposes before authorizing it for consequential actions, and for keeping the scope of each agent within what you actually want it to do. This applies whether an action is requested in chat, through an in-chat form or control, from the CLI, by an SDK/API application, or by an MCP client.
q5m provides software-generated assistance that can be inaccurate. It is not a substitute for professional advice or emergency services. Health, financial, legal, and other domain-specific guidance should be treated as a starting point, not as one-on-one counsel.
3. Third-party integrations
The service integrates with third-party platforms such as Google, DeepSeek, Anthropic, OpenAI, Intervals.icu, Monarch, Home Assistant, and others. Third-party downtime, API changes, terms of service updates, data practices, or account issues on those platforms can affect agent behavior outside our control. q5m Basic sends prompts and relevant agent context to DeepSeek. By enabling q5m Basic, you agree to DeepSeek's Terms of Use.
4. Access methods and inference
In the q5m web chat, q5m runs the agent conversation and tool calls using q5m Basic or a model provider you connect (bring your own key, or BYOK). Included inference is subject to the current allowance. BYOK use is subject to your provider's prices, limits, and terms. In-chat interfaces can display results and let you submit data or request actions; they are part of the Service, not a separate guarantee that an output or action is correct.
The CLI and MCP let terminal users, scripts, and compatible clients access q5m tools, instructions, and saved data. For client-driven workflows, the client or script controls orchestration and any model inference it uses; these tool connections do not themselves provide a q5m-hosted chat model. Any third-party client or model charges are governed by your arrangement with that provider, not automatically included in q5m Basic.
The SDK and APIs let applications authenticate users and access supported q5m data and services. Available capabilities depend on the endpoint, your permissions, and configuration. Direct data and tool endpoints do not themselves provide chat inference. Chat APIs, including the OpenAI-compatible chat endpoint, run q5m-hosted agent inference using your configured model access, subject to applicable allowances or provider charges and terms. Third-party clients and applications have their own terms and data practices.
5. Privacy
How account, profile, conversation, and integration data are handled is described in our Privacy Policy.
6. Service warranty
To the fullest extent permitted by law, the service is provided "as is" during this early phase, without warranties of uninterrupted operation, fitness for a specific purpose, or guaranteed outcomes from any agent.
7. Changes
We may update these Terms over time. Continued use after updates means you accept the revised Terms.
8. Contact
Questions can be sent to [email protected].