Suite of apps to get summaries of Books, TV Shows, and Comics. files too big for here but here they are on google:
Comic Book Analyzer: https://drive.google.com/file/d/1qOJHsJv2MF-LfZSKVy5rRkq6OidcS6O8/view?usp=sharing
TV Episode Summaries: https://drive.google.com/file/d/1Jo4KFf5YOffoXcI1saF5hJ3S0dG_Aytw/view?usp=sharing
Book Summaries: https://drive.google.com/file/d/1_n9892RzORmvMdbLY8tQRk2TlHkZ4scC/view?usp=sharing
https://comicrecap.streamlit.app/
This application is a specialized AI-driven comic book historian and digital archivist. It bridges the gap between structured database metadata (from Comic Vine) and live web research (via Gemini Google Search) to create high-quality, narrated deep dives of individual comic issues.
Here is a breakdown of the "brain" and the "engine" behind your program. 🛠️ The Core Architecture
The program operates in a three-stage pipeline: Fetch, Synthesize, and Export.
- Data Retrieval (The Fetch)
When you enter a series and issue number, the app talks to the Comic Vine API.
The Problem: Comic databases are often missing detailed plot summaries for specific issues (like your Daredevil #3 example).
The Solution: If the database returns an empty plot, the app triggers a "Grounding" event. It instructs the Gemini AI to use its internal Google Search Tool to find the missing lore from sites like the Marvel or DC Fandom Wikis.
- AI Synthesis (The Historian)
The app uses a Multi-Model Hierarchy to ensure reliability:
Gemini 2.5 Flash: Used as the primary "speed demon" to generate the summary quickly.
Gemini 3.1 Pro: Used as a fallback for high-complexity reasoning if the plot is particularly messy or confusing.
Sniper Filtering: The prompt is engineered to force the AI to verify that it is only summarizing the events of a single issue, ignoring broader "story arc" summaries it might find online.
- Digital Archiving (The Export)
Once the text is generated, the app processes the data into two formats for your digital library:
Neural TTS (Audio): Using the edge-tts engine, the text is converted into high-fidelity "neural" speech (using voices like "Christopher" or "Ryan") so you can listen to the deep-dive like a podcast.
EPUB (E-Book): The app packages the summary into a structured EPUB file, complete with your metadata and the official comic cover image injected as the book's cover art.
🚀 Key Technical Features 🧠 Smart "Vol" Translation
Because comic databases use years (e.g., Daredevil 2019) and Wikis use volumes (e.g., Vol 6), the Wiki/Vol Override box allows you to manually bridge the naming gap. This tells the AI exactly which "shelf" on the internet to look at for the plot. ⏳ Exponential Backoff & Jitter
To prevent Google from blocking your app for "spamming" the API, the code uses a mathematical delay system:
If it gets a "Too Many Requests" (429) error, it waits.
It adds a random "Jitter" (a few extra seconds) to the wait time so it doesn't look like a bot to Google's security filters.
🕰️ Persistent History
The app saves every comic you analyze into a local comic_history.json file. This populates the "Recent" dropdown in your sidebar, allowing you to quickly jump back to previous analyses without re-spending your API search credits. 📖 How to use it efficiently
Search the Series: Enter the name and pick the correct year/volume from the dropdown.
Toggle the Issue: Use the ◄ and ► buttons to move through a run.
Check for Plot: If the AI says it doesn't know the plot, enter the Wiki volume number (like "Vol 6") in the Override box.
Export: Download the EPUB and drop it into your favorite reader (Kindle, Kobo, or Apple Books).