Tools

Every model, more capable

Babbily reads your videos, recordings, PDFs and spreadsheets for whichever model you pick, even one that can’t open that file on its own. Search, images, canvas and memory come built in.

Babbily reads your videos, recordings, PDFs and spreadsheets for whichever model you pick, even one that can’t open that file on its own. Search, images, canvas and memory come built in.

Babbily reads your videos, recordings, PDFs and spreadsheets for whichever model you pick, even one that can’t open that file on its own. Search, images, canvas and memory come built in.

Try it in Babbily

Works with every chat model in Babbily. Babbily doesn’t train on your chats by default.

Kickoff callShare
Kickoff call.mp4Video · 14:32
What did we decide about the launch date, and who owns what?

You moved the launch to March 4, so the beta can run for two full weeks.

What was decided
  • 06:12 — Maya proposes March 4 instead of February 18.
  • 09:40 — Jun says the beta needs two full weeks of feedback.
  • 12:05 — The launch slide on screen changes to “Launch · Mar 4”.
Who owns what
  • Maya: the launch plan and announcement draft
  • Jun: a summary of beta feedback by February 25
  • Alex: the pricing page update
Claude Opus 5.5 can make mistakes. Please double-check responses.

Kickoff call.mp4

Video · read for Claude

Kickoff callShare
Kickoff call.mp4Video · 14:32
What did we decide about the launch date, and who owns what?

You moved the launch to March 4, so the beta can run for two full weeks.

What was decided
  • 06:12 — Maya proposes March 4 instead of February 18.
  • 09:40 — Jun says the beta needs two full weeks of feedback.
  • 12:05 — The launch slide on screen changes to “Launch · Mar 4”.
Who owns what
  • Maya: the launch plan and announcement draft
  • Jun: a summary of beta feedback by February 25
  • Alex: the pricing page update
Claude Opus 5.5 can make mistakes. Please double-check responses.

Kickoff call.mp4

Video · read for Claude

01

Files any model can read

Video, audio, PDFs and Office files, read before the model answers.

Jump to

01

Files any model can read

Video, audio, PDFs and Office files, read before the model answers.

Jump to

02

Video, with context

A transcript, scenes and on-screen text, all with timestamps.

Jump to

02

Video, with context

A transcript, scenes and on-screen text, all with timestamps.

Jump to

03

Tools in every chat

Search, research, images, canvas, memory and your apps.

Jump to

03

Tools in every chat

Search, research, images, canvas, memory and your apps.

Jump to

How Babbily reads your files

Attach a file and ask. Babbily turns it into something the model can use.

Attach anything

Video, audio, PDFs, Office files, images, code or a ZIP of several files. Pick any model you like.

Babbily reads it

Video and audio become a transcript with timestamps and speakers. PDFs keep their headings and tables.

The model answers

The model you picked works from what Babbily found, even if it can’t open that kind of file itself.

Read once

Follow-up questions, edits and retries reuse the first read, so the same file isn’t read twice.

Under the hood

A model that can’t watch video can still answer about one

When a model has no video input, Babbily watches for it. It writes a transcript with timestamps and speaker labels, describes each scene and reads the text on screen. The model answers from that, so it can point to the moment that matters.

Video and audio

Transcripts with timestamps and speakers.

PDFs

Headings kept, tables rebuilt, figures described.

Images

Described, with their text, for text-only models.

A video goes to a model that can't watch it: Babbily reads it first, then the model answers with timestamps.

Frequently asked questions

Bring any file to any model

Attach a video, a PDF or a spreadsheet and ask the model you like best.

Try Babbily free

Read the docs ↗