Product context, available when your AI needs it.
This guide covers the normal setup and the small number of concepts worth knowing. After installation, most work happens naturally inside your coding-agent conversation.
Install Vicaura
Paste the setup prompt from the agent install page into Codex or Claude Code. The agent runs the installer, registers Vicaura’s MCP server, installs its usage skill, and opens secure sign-in.
npx --yes @vicaura/agent@latest install
# If npx is unavailable:
pnpm dlx @vicaura/agent@latest installInstallation changes only the selected coding agent’s Vicaura skill and MCP entry. It does not rewrite unrelated configuration.
What “automatic” means
The Vicaura skill teaches your AI when product intelligence can materially improve a task. For product strategy, positioning, pricing, messaging, customer, feature, and go-to-market questions, the agent can look for relevant context without you naming Vicaura.
It first checks whether your saved products fit the product you are building. If they do not, it can start a better-matched search, wait for the research to complete, generate the files the task needs, and then answer with that evidence.
Ask in your own words
“Help me identify the ICP and category position for a LinkedIn prospecting SaaS.”
“What pricing and packaging would make sense for this product?”
“Recommend the next three features and explain which relevant products informed the reasoning.”
“Use Vicaura to find products like this and create a Buyer Persona file.”
You can ask naturally or name Vicaura explicitly. Product examples are reference context, not assumptions about your own product.
Repos, files, and usage
Using Vicaura inside an AI agent requires an active monthly subscription. Agent-started searches and file generations use the same account-wide credits shown on the website.
Each analyzed product becomes a repo of reusable product context. Reading an existing repo does not create duplicate files. When a task needs missing evidence, the agent may use a product repo discovery or generation credits just as the website would. Longer actions expose progress and wait for usable results before the final answer.
Your AI can retrieve multiple relevant products and files. It should compare them, cite product influence naturally where useful, and keep doing its own reasoning rather than repeating Vicaura content verbatim.
Check the connection
Run the doctor command when Vicaura is not appearing, after updating your coding agent, or after changing MCP configuration.
npx --yes @vicaura/agent@latest doctor
# If npx is unavailable:
pnpm dlx @vicaura/agent@latest doctorIf the check reports an authentication issue, reconnect Vicaura from the MCP settings in your coding agent. To install the current skill and integration again, run:
npx --yes @vicaura/agent@latest update
# If npx is unavailable:
pnpm dlx @vicaura/agent@latest updateCopy one setup prompt and let your coding agent handle the rest.
Open agent install