atypica Research Skill
A skill for atypica.ai studies: AI-simulated consumers answer questions and the agent compiles a report with sources.
- 62 ★ on the repository
- Last commit
Extended access
Why this level
- Requests and answers go to an external service and a study spends credits of a paid account.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/atypica-research-skillDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Requests and answers go to an external service and a study spends credits of a paid account.
This is third-party code. Review the repository files before installing.
What it does
The skill teaches an agent to run research through the atypica.ai MCP server. The user states a question about consumers, a product concept or a market, confirms the plan, and the service interviews AI-simulated consumers who discuss the question. The output is a report with traceable sources and, if needed, a podcast. Server tools cover study creation, sending messages, getting steps and messages, listing studies, the report, the podcast and panel and persona search. Without an MCP client any tool is one request through `scripts/mcp-call.sh`. A study runs from half an hour to an hour and the skill asks the agent to keep the user informed. Results must be presented as an AI simulation.
Who it is for. Product managers and marketers who test consumer hypotheses before live interviews and want a fast draft report.
Good fit when
- The audience reaction to a product concept must be tested.
- Market research needs a report with sources within an hour or two.
- A hypothesis should be run across different personas and panels.
- A podcast from the study results is needed.
Not a fit when
- Answers from real people are needed, not a simulation.
- There is no atypica.ai account and key.
- The study must finish in minutes.
Example request
Run a study: how students would react to a 990 ruble monthly online course subscription, and show the report with sources.Limitations
An atypica.ai account and an API key of the `atypica_xxx` form are needed. A study spends account credits, and usage is visible on the tokens page. Answers come from a simulation of AI consumers, not a live sample, so conclusions should be confirmed with real interviews. Request data goes to an external service. No licence is stated in the repository. The skill connects over HTTP with an `Authorization` header, so keep the key in a secure store rather than shared configuration.
How to disable. Remove the `atypica-research` entry from the MCP client config, revoke the key on the atypica.ai keys page and delete the skill folder.
FAQ
Where do I get a key?
On `atypica.ai/account/api-keys`, in the `atypica_xxx` format.
How do I connect without an MCP client?
Export `ATYPICA_TOKEN` and call tools through `scripts/mcp-call.sh`, for example `atypica_study_create`.
How long does a study take?
Usually half an hour to an hour after plan confirmation. The skill asks the agent to keep the user informed.
What results can I get?
A report with traceable sources via `atypica_study_get_report` and a podcast via `atypica_study_get_podcast`.
Whose answers are these?
An AI simulation of consumers. The skill requires results to be presented that way.
Discussion
No questions or reviews yet.
Related
SkillsA skill that runs user research inside the agent: plan, interview guide, report synthesis, plus interviews and quantitative surveys through the Cookiy AI servic
The official remote MCP server for the WebAsk survey builder plus 25 skills: surveys from a brief, pre-launch checks, answers and reports
Skill search from inside the conversation: the agent finds a SKILL.md, reads it and applies it with no prior install
Skills for researching Reddit from an agent: semantic search over posts and comments, trends, subreddit discovery and lead finding via reddapi.dev
An autonomous deep-research agent: gathers sources, verifies them and writes a report on any question
Claude Skills MCP: semantic search for skills
Claude Skills MCP
An MCP server with vector search: the agent describes a task and gets matching skills from the Anthropic and K-Dense collections
README in short
A skill for research with AI-simulated consumers through the atypica.ai MCP: from question and plan to a sourced report and podcast.