Flashback
an agent skill that references 127 years of design trends
Last verified:
What is Flashback?
Flashback is a command-line knowledge store designed for developers who want a fast, local, scriptable memory system. It captures text, URLs, and commands from your workflow, extracts structured metadata automatically, and makes everything searchable without requiring cloud services.
Key features include a CLI-first workflow built with Cobra for seamless integration into developer workflows, metadata extraction for URLs using OpenGraph, Twitter cards, and JSON-LD, and AI enrichment powered by Google Gemini with strict JSON schema enforcement for deterministic results. The tool uses local-first storage backed by Turso/libSQL, includes a TUI (Terminal User Interface) viewer built with Bubbletea for browsing entries, offers fast search with filters and tags, and maintains a separate metadata table for incremental enrichment without schema churn.
Flashback processes inputs through a simple pipeline: it detects the type (text, URL, or code), scrapes metadata, runs Gemini enrichment to add tags and summaries, and stores everything in SQLite/Turso with structured metadata. It is specifically designed for developers who need to store and retrieve code snippets, commands, documentation URLs, and technical notes quickly from their local machine.
The tool is installable via Homebrew, from source using git and make, or directly with go install. Configuration is stored at ~/.config/flashback/config.yaml, and the data model separates flashbacks from metadata to allow multiple enrichment passes.
The roadmap includes adding more site-specific extractors for GitHub, YouTube, Reddit, and Medium, configurable metadata schema, and a plugin system for custom enrichers.
Flashback pricing
Pricing model: Freemium
Flashback is free and open-source software. It is hosted on GitHub as a public repository with no paid tiers mentioned. The tool requires a Google Gemini API key for AI enrichment features, which means users must pay for Gemini API usage separately based on Google's pricing. Storage uses Turso/libSQL which has its own pricing if using Turso's cloud service, but local SQLite storage is free.
Flashback pros
- Command-line first workflow for seamless developer integration
- Local-first storage - no cloud dependency required
- Fast semantic search with filters and tags
- Metadata extraction from URLs using OpenGraph, Twitter, JSON-LD
- AI enrichment powered by Google Gemini
- Strict JSON schema enforcement for deterministic AI output
- TUI viewer built with Bubbletea for browsing entries
- Separate metadata table allows incremental enrichment
- Scriptable through shell commands
- Built with Go for fast performance
- Configurable via YAML config file
- Homebrew install support for easy setup
- Detects input type automatically (text, URL, code)
- Tags and summaries generated automatically
- SQLite/Turso backend for reliable storage
Flashback cons
- Only one contributor (yagnikpt) - single maintainer risk
- Very low adoption - only 12 GitHub stars
- No plugin system yet (planned for roadmap)
- Limited site-specific extractors currently
- No configurable metadata schema yet (planned)
- Requires Google Gemini API key for AI features
- Browser-based URL metadata may fail for some sites
- TUI is basic compared to full GUI applications
- No web interface - CLI/TUI only
- Recent project - may have unstabilized features
- Limited documentation beyond README
- No enterprise features or multi-user support
- Windows support not explicitly mentioned
- Active development - possible breaking changes
- No mobile app for viewing entries
Frequently asked questions about Flashback
What is Flashback?
Flashback is a command-line knowledge store that captures text, URLs, and commands, extracts structured metadata, and makes everything searchable. It is designed for developers who want a fast, local, scriptable memory system.
How do I install Flashback?
You can install Flashback using Homebrew with 'brew install yagnikpt/tap/flashback', from source by cloning the repository and running 'make install', or directly with 'go install github.com/yagnikpt/flashback@latest'.
How do I add entries to Flashback?
Use the 'flashback add' command. You can add commands like 'flashback add kubectl rollout restart deployment web' or URLs like 'flashback add https://blog.bytebytego.com/p/understanding-load-balancers'.
How do I search in Flashback?
Use the 'flashback search' command with your search terms. For example: 'flashback search load balancer' or 'flashback search kubernetes'.
What metadata does Flashback extract from URLs?
Flashback extracts metadata using OpenGraph, Twitter cards, and JSON-LD. Example metadata fields include tldr, description, tags (as a JSON array), and image (URL).
How does the AI enrichment work?
Flashback uses Google Gemini for AI enrichment with strict JSON schema enforcement to ensure deterministic results. It automatically generates tags and summaries for your entries.
Where is Flashback data stored?
Flashback uses local-first storage backed by Turso/libSQL. Records are stored in the 'flashbacks' table with metadata stored separately in the 'metadata' table as key/value pairs.
Where is the configuration file located?
The configuration file is located at ~/.config/flashback/config.yaml.
What is the TUI in Flashback?
The TUI (Terminal User Interface) viewer is built with Bubbletea. You can launch it by running 'flashback' without any arguments to browse your entries.
What is planned for the Flashback roadmap?
The roadmap includes more site-specific extractors for GitHub, YouTube, Reddit, and Medium, a configurable metadata schema, and a plugin system for custom enrichers.