How Snowman and the web dashboard work
Snowman’s web dashboard is the single place where you talk to the AI assistant, design scenarios, run scheduled tasks, view analytics, and set up API integrations. Below is an overview of the key sections and how they connect together.
If you’re just getting started, read this page end-to-end and then go to “Snowman API”to connect your services via HTTP.
Core modules: what Snowman is made of
The flow is simple: you describe a task in plain language in chat, then turn it into a formal scenario and, if needed, schedule it in the Scheduler. Everything lives inside spaces (projects), and results and limits are visible in analytics. For external systems, use API keys.
Chats with Snowman
Chat is the entry point. Describe tasks in everyday language: analyze, collect, compare, generate ideas. Snowman helps clarify the goal, suggests a report format, and can recommend how to package the task as a scenario.
- Quickly validate hypotheses and phrasing for future scenarios.
- Discuss task logic and the expected output.
- Save a good approach as a reusable scenario.
Spaces and teams
A space is a separate “world” for a project, department, or client. It contains its own scenarios, scheduler jobs, chat context, and limits. On team plans, a space can belong to an entire team.
- Separate marketing, sales, analytics, and client work.
- Share access to spaces with teammates on team plans.
- Keep sensitive scenarios in dedicated spaces.
Task Scheduler
The Scheduler runs scenarios regularly without your manual involvement. You describe what needs to be done once, choose a schedule, and Snowman takes care of the queue and stores the results.
- Daily/weekly reports, monitoring, storefront checks.
- Monthly schedules or cron-like patterns.
- Manual run option to override the schedule.
Analytics and usage
Analytics shows how Snowman is used: which scenarios run most often, how many scheduler runs happen per month, and how much quota you have left on your plan.
- Summary of scenario runs and consumed resources.
- Per-job run history for Scheduler tasks.
- Helps choose the right plan for your real workload.
Records and skills: teaching Snowman new moves
Records (“show Snowman by doing”)
In Records, you can “show” Snowman how you do a task in the browser: navigation, clicks, forms. Sessions are saved as recordings that Snowman can later use to build skills.
- Record a typical manual operation once.
- Use a record as the basis for a future scenario.
- Transfer expert knowledge to Snowman without a long spec.
Skills
A skill is a generalized “technique” Snowman can reuse across scenarios. Skills are built from records and fine-tuning: you explain what the skill does and which parameters it needs.
- Extract reusable sub-tasks (e.g., “collect product cards”).
- Reuse the same skills in different scenarios.
- Over time, Snowman becomes closer to your business context.
API keys and external integrations
How API keys work
When the web dashboard is not enough, use the API. In “API access” you can issue three key types: Tenant, Scenario, and Read-only. Each covers a different scope: general access, triggering a specific scenario, or reading history only.
The full technical spec and ready-to-use curl examples are in the dedicated docs: “Snowman API”. There you can also see plan-based limits.
- Tenant keys — for backend services and general integrations.
- Scenario keys — for triggering a specific Scheduler scenario.
- Read-only — for safely reading history and exporting results.
Security
Snowman runs tasks in your environment: in your browser, with your sessions, and within your access rights. The dashboard does not store website passwords, and API keys are issued per scenario/integration so access is easy to manage.
- Revoke API keys in one click without breaking other integrations.
- Scenarios run within the limits you define (domains, frequency, quotas).
- For BI/reporting, use read-only keys without run permissions.