Community Ratings & Leaderboards

How the Voice Ranking System Works

The voice ranking system is a sophisticated, autonomous engine designed to bring clarity and precision to voice discovery. It's built on a unique hybrid architecture that combines intelligent AI analysis with the power of real user feedback.

The system operates in two distinct but connected ways: by providing immediate, contextual rankings for any given script, and by maintaining dynamic, global leaderboards that reflect the most effective voices over time.

Part 1: Finding the "Perfect Match" (The Per-Script Ranking)

When a script is submitted for analysis, the system's primary goal is to generate the most contextually relevant voice recommendations for that specific text. This process involves several automated steps.

Step 1: The Casting Director AI

A proprietary Casting Director AI analyzes any submitted script. It deconstructs the text to understand its core intent based on several attributes, including but limited to:

  • Tone & Mood: Is the content serious, playful, inspirational, or calm?
  • Pacing & Energy: Does it require a fast, energetic delivery or a slow, deliberate narration?
  • Technical Details: What is the ideal pitch, intensity, and speaking rate (WPM)?
  • Application: Is this a script for a documentary, a customer support line, an audiobook, or a movie trailer?

Step 2: Generating the "Ideal Voice Profile"

From this analysis, an "Ideal Voice Profile" is generated. This serves as a detailed, data-driven blueprint of the perfect voice for the submitted script.

Step 3: Consulting the Community Score

Before voices are compared, the system consults the Community Score. This score is built from the 1-5 star ratings that users have given to voices for specific styles. A voice with a high average rating for Educational content receives a significant score bonus when being considered for an educational script. This ensures that voices already validated by the community are given greater weight.

Step 4: The Competitive Ranking

The system then scores every voice in the library against the "Ideal Voice Profile." The final ranking is determined by:

  • The technical similarity of the voice to the ideal profile (pitch, WPM, etc.).
  • The stylistic match between the voice's tagged styles (Dramatic, Calm) and the styles required by the script.
  • The Community Score bonus the voice has earned for those specific styles.

The result is an instant, highly accurate list of the best voices for a specific project, with recommendations backed by both AI precision and real-world user satisfaction.

Part 2: The Global Leaderboards (e.g., "Top 100 Dramatic Voices")

The global leaderboards are an emergent property—a "side effect"—of all the individual "Perfect Match" rankings happening on the platform. They are designed to be dynamic, fair, and always reflect what is currently effective.

Step 1: Awarding Leaderboard Points

Each "Perfect Match" ranking is considered a competitive event. The top 100 voices from that ranking earn Leaderboard Points, with higher-ranking voices receiving more points. These points are then applied to each relevant category leaderboard. For example, if a voice ranked #1 for a dramatic narration script, it would receive a large number of points on both the Dramatic and Narration leaderboards.

Step 2: Relevance Over Time (The Decay Model)

User preferences and TTS technology evolve. To keep the leaderboards fresh and fair, the system incorporates a decay model.

Think of a voice's score like a "heat score":

  • When a voice earns points, its score gets "hotter."
  • If a voice doesn't consistently perform well in new competitive events, its score slowly "cools down" over time.

This decay ensures that the leaderboards reflect current relevance, giving new, high-quality voices a fair chance to rise to the top and preventing the lists from becoming stale.

The leaderboards are a living, breathing reflection of the best voices on the platform right now, making it easy to discover the top-performing and most popular choices for any style.

The Community's Role: The Power of Feedback

User ratings are the most valuable signal within the ecosystem, directly influencing the intelligence of the ranking engine.

When a user rates a voice for a specific text (e.g., giving 5 stars to a voice for an Educational script), that rating updates the voice's Community Score for the Educational category.

This score then acts as a powerful bonus (or penalty) the next time that voice is considered for an educational script. This creates a powerful feedback loop where voices consistently validated by the community naturally rise to the top in future rankings, making the entire system smarter and more accurate for everyone.