Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    voxprice voice pricing workflow

    1

    Build and manage voice-to-quotation workflows using FastAPI, Google Gemini, and PostgreSQL.

    Free

    Security scanned
    voxprice voice pricing workflow

    voxprice voice pricing workflow

    Example session with this skill installed

    I need to set up the product matching workflow. How should I handle items that Gemini extracts but aren't in my PostgreSQL catalog?

    • Read your context and instructions
    • Compiled the voxprice voice pricing

    When an item is extracted from audio but doesn't have a confident match in your catalog, the workflow must prevent automated guessing to avoid pricing errors.

    1. Extraction as Draft: The system labels the Gemini output as unverified draft data.
    2. Manual Flagging: Instead of forcing a match, the UI flags the line item for manual selection.
    3. Human Review: The operator must select the correct product from the Admin-configured catalog.

    Server Validation: The final pricing is only calculated once the product_id corresponds to a valid entry in the database.

    Next steps

    • Ensure your product catalog descriptions are detailed enough for Gemini to provide better initial candidates.
    • Review the Admin/CMS settings to adjust the extraction prompt if generic product names are frequently failing.
    • Check backend logs to see the confidence score of the failed matches.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Voice-to-quotation workflows are prone to transcription errors and hallucinated pricing when handled by generic LLM prompts. This skill provides the logic to build and manage VoxPrice, a FastAPI and vanilla JavaScript application that uses Google Gemini for structured audio extraction and PostgreSQL for authoritative server-side pricing.

    What it does

    • Structured Audio Extraction helps extract product names and quantities from audio files or microphone recordings via Google Gemini.
    • Human-in-the-Loop Review enforces a verification step where extracted data is treated as unverified drafts until a human confirms.
    • Server-Side Pricing ignores browser-calculated totals in favor of secure backend calculations against a product catalog.
    • Admin CMS Management configures product catalogs, landing pages, Gemini model parameters, and role-based access control.
    • Persistence Management handles quotation history and session-based authentication via SQLAlchingy and PostgreSQL.

    How it works

    1. Source Audio: Record via browser mic (WAV) or upload an existing file.
    2. Extract Drafts: Route audio through the FastAPI backend to Gemini for structured detail extraction.
    3. Verify Items: Inspect and correct extracted product names and quantities against the active catalog.
    4. Calculate and Commit: Request a server-side recalculation and save the finalized quotation to the database.

    Frameworks & tools

    Python, FastAPI, SQLAlchemy, PostgreSQL, Google Gemini API, and vanilla JavaScript.

    Why this beats prompting it yourself

    Prompting alone often ignores business logic and database integrity. This skill separates LLM-drafted extraction from server-side price truth, ensuring you never save a quotation based on a hallucinated total.

    Use cases

    • Build a voice-ordered inventory system for warehouse operators.
    • Implement a sales tool that converts voice memos into formal quotes.
    • Configure role-based access for agents and managers in a pricing workflow.
    • Troubleshoot Gemini extraction precision by refining catalog matching logic.

    Known limitations

    Does not independently host the app, provide Gemini API keys, or access live microphones. Requires a persistent Python environment; static hosting is insufficient.

    Requires Gemini API key and Python/FastAPI environment. Not compatible with static-only hosting. Does not access microphone or DB directly without the deployed VoxPrice application.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 6 days ago

    • Free with an Agensi account

    Permissions

    Read Files
    Network Access
    Listed6 days ago

    What's inside

    Frequently Asked Questions

    VoxPrice prevents "pricing hallucinations" by using Gemini solely for data extraction, while forcing all final calculations to occur on the server against your secure PostgreSQL product catalog.

    The package includes the full FastAPI backend logic, the vanilla JavaScript frontend, SQLAlchemy database schemas, and the administrative CMS for managing your product list and model parameters.

    This skill requires a persistent Python environment to run the FastAPI server and a PostgreSQL database; it is not compatible with static-only hosting services like GitHub Pages.

    Yes, the built-in Admin CMS allows you to adjust Gemini model parameters and refine catalog matching logic to improve extraction precision for your specific industry terms.

    A 'Human-in-the-Loop' step is integrated, meaning all AI-extracted data is saved as an unverified draft that must be reviewed and confirmed by a user before a final quotation is generated.

    New in APIs & Backend