Skip to content

speakr

v0.9.3-alpha Security

This release includes 1 security fix for security teams reviewing exposed deployments.

Published 1mo Productivity & Wikis
✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →
This release patches 1 known CVE

Affected surfaces

rce_ssrf deps breaking_upgrade

Summary

AI summary

Patches CVE-2026-8461 by updating bundled FFmpeg to version 8.1.2 from BtbN/FFmpeg-Builds.

Full changelog

Release Notes - v0.9.3-alpha

A security patch that updates the bundled FFmpeg and changes where Speakr sources it.

Security

  • Patches CVE-2026-8461 ("PixelSmash"). Speakr runs ffprobe and ffmpeg on uploaded media for codec detection, conversion, and audio extraction. The FFmpeg that earlier images bundled (johnvansickle static 7.0.2) contained a heap out-of-bounds write in the MagicYUV decoder that a crafted media file could use to crash the worker or, in some cases, achieve remote code execution. The flaw is fixed in FFmpeg 8.1.2. Because the johnvansickle static builds have been frozen at 7.0.2 since 2024, the Docker image now sources FFmpeg from the actively maintained BtbN/FFmpeg-Builds project, pinned to the 8.1 release branch (currently 8.1.2). Multi-user instances that accept uploads from untrusted users were the most exposed.

Compatibility

No configuration or data changes. Upgrade by pulling the new image with docker compose pull && docker compose up -d. This upgrade is recommended for all deployments, and especially for any instance that accepts uploads from users you do not fully trust.

Security Fixes

  • CVE-2026-8461 — heap out-of-bounds write in MagicYUV decoder fixed by updating FFmpeg to version 8.1.2 (BtbN/FFmpeg-Builds)

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Track speakr

Get notified when new releases ship.

Sign up free

About speakr

Speakr is a personal, self-hosted web application designed for transcribing audio recordings

All releases →

Related context

Earlier breaking changes

  • v0.10.0-alpha Unauthenticated API requests now return a JSON 401 instead of a 302 redirect.

Beta — feedback welcome: [email protected]