Open-source intelligence research console

SignalGlass

Follow signals.
Keep custody of your research.

SignalGlass is a local-first research console built for journalists, investigators, researchers and curious minds who need to move through public information without losing context, sources or control of the workflow.

born local yt-dlp SearXNG mpv SQLite
research session
signalglass › search "skills" --channel @creator
scanning channel sources / resolving metadata
01 / matched signal
source: youtube · metadata resolved · public source
02 / related channel result
title · date · views · channel · canonical URL
03 / broader discovery pivot
SearXNG → web / images / public-source discovery
signalglass › _
01 / Origin

It began with a simple frustration.

01

The information was there.

Useful material lived across videos, channels, search engines, metadata and source pages—but the research workflow was fragmented across too many tabs and too many tools.

02

The workflow was missing.

We wanted one place to search deeply, inspect what we found, organize sources, revisit previous work and move immediately from discovery to examination.

03

So the interface grew around the work.

yt-dlp became the lens into channels and video metadata. SearXNG opened the wider web. Collections created continuity. mpv made playback immediate.

04

SignalGlass emerged.

Not another feed. Not another hosted workspace. A focused, operator-controlled console for following public signals through noise.

Search is easy. Investigation is what happens after the first result.
02 / Architecture

A browser interface. A local operational core.

Researcher
query
investigation
judgment
SignalGlass
local console
workflow
orchestration
Discovery
yt-dlp
SearXNG
public sources
Evidence context
metadata
history
collections
Local tools
mpv
SQLite
filesystem
03 / Capabilities

Built around the questions researchers actually ask.

[ SEARCH ]

Channel research

Examine recent uploads or search deeply within a creator's publication history.

[ TRACE ]

Metadata inspection

Resolve titles, dates, channels, view counts and canonical source URLs for faster examination.

[ PIVOT ]

Open-web discovery

Use SearXNG to pivot beyond one platform into broader web, image and source discovery.

[ ORGANIZE ]

Collections

Group channels and sources around investigations instead of rebuilding context every session.

[ RECALL ]

Search history

Preserve a local record of previous research activity and results for continuity across sessions.

[ EXAMINE ]

Local playback

Send discovered video directly to mpv without forcing the research workflow back into a platform interface.

04 / The instrument

This isn't a concept.
It's already running.

SignalGlass began as a working research tool before it had a name. The interface follows the investigation: search, narrow, inspect, organize and open the source.

SIGNALGLASS // LOCAL RESEARCH CONSOLE
+ SignalGlass running locally Product image / Day Zero build
LOCAL-FIRST PUBLIC SOURCES OPERATOR CONTROLLED WORKING SOFTWARE
05 / Operating principles

Research tooling should respect the operator.

Local-first

SignalGlass is designed around an application running on the researcher's own machine, not a mandatory hosted workspace.

Source-oriented

The goal is not to manufacture answers. It is to help the operator find, inspect and move between public sources efficiently.

Composable

Proven command-line tools remain useful underneath the interface rather than being hidden behind an opaque proprietary stack.

Human judgment stays upstream

Search and automation reduce friction. They do not replace source evaluation, verification or editorial judgment.

SignalGlass / Day Zero

See the signal.
Preserve the trail.

SignalGlass is being built as a practical research instrument: discover the source, examine the context, preserve the path, and keep the operator in control.

DAY ZERO // SIGNALGLASS.ORG // 2026