Skip to content
SPECIFICTION
Get listed

Claras vs YouTubeTLDR

Claras

meetclaras.com

YouTubeTLDR

github.com

Side by side

FigureClarasYouTubeTLDR
Price from$4.39/mo—
Free planYes—
Free trial——
Open source——
WebYesYes
Windows—Yes
macOS—Yes
Linux—Yes
Self-hosted—Yes
Plans published40
Spec score · AI Video Summarizers8.4 (#7)6.9 (#25)

Plans

Claras

  • FreeFree
  • Pro$4.39 a month
  • Unlimited$6 a month
  • Lifetime$179, paid once

Claras pricing →

YouTubeTLDR

No plans published.

YouTubeTLDR pricing →

In their own words

Claras

Claras is a web service and Chrome extension for transcribing YouTube videos and working with their transcripts. It creates AI summaries, clickable chapters, timestamps and transcript-based chat, and SuperSearch can summarize a playlist or YouTube search page in one go. It supports 17 languages, with automatic language detection and summaries in a preferred language when captions are available. Supported inputs include YouTube watch and youtu.be links, Shorts, live videos and embeds. Transcripts can be sent to Claude, ChatGPT, Gemini or NotebookLM, or exported as TXT, PDF or.srt files. The free plan includes 3 videos and 1 SuperSearch per month; Pro is $4.39 a month, and Unlimited is $6 a month. Unlimited and Lifetime users can use their own API key for models including ChatGPT, Claude and Gemini. Claras says transcript text is not stored on its servers, though it is sent to third-party AI providers for summaries and chat.

Visit meetclaras.com

YouTubeTLDR

YouTubeTLDR is a free, self-hosted YouTube video summarizer that uses Gemini AI. Users provide a YouTube video URL, choose an available Gemini model and customize the AI’s instructions to shape the summary. The app also provides the full raw transcript, and saves summaries locally in the browser for later reference. Its Rust server runs on the user’s machine; users supply a Google Gemini API key, which the project says is used directly without proxies or data collection. The project describes the application as a roughly 0.3 MB single binary, with no database, Tokio or frameworks, and a vanilla HTML and JavaScript interface. The documented setup is to run the executable and open localhost:8000 in a browser. Server IP, port and worker count can be changed using environment variables. It supports Linux, macOS, Windows, web and self-hosted use, and the repository identifies the license as MIT. The project says it is optimized for personal use, not hundreds of concurrent users. It does not support video uploads or chapter generation.

Visit github.com