Darkfish Foundry v0.8.1.2

What is Darkfish Foundry

Darkfish Foundry: Local-First Artifact Generation for Human and Agent-Composable Workflows

Most individuals and small teams already have valuable knowledge. It is just trapped in the wrong forms: video transcripts, notes, Markdown files, text files, JSON/YAML/CSV/log documents, PDFs, internal records, and scattered local source material. Darkfish Foundry is built to help turn that material into structured, inspectable, reusable artifacts without requiring a cloud document-processing service or an enterprise data team.

Darkfish Foundry v0.8.1.2 is a local-first, privacy-first source intake and artifact-generation application for YouTube transcript jobs, local document processing, and native-text PDF extraction. It provides both a Tauri desktop GUI for human users and an agent-facing CLI/STDIO surface for local automation, scripts, and agent harnesses.

The project’s core purpose is to give individuals, researchers, builders, agents, and small businesses a practical entry point for processing their own source data. Darkfish converts raw materials into local files, manifests, metadata, machine-readable outputs, and structured artifacts that can support review, archiving, automation, training-data preparation, and human-agent workflows.

A key design goal is composability. Darkfish is not only an application that performs a fixed set of processing tasks; it is also a local toolbox with an agent-readable action dictionary. That means an authorized local agent, developer, or machine-human team can inspect the available commands, understand the artifact structure, and iteratively adapt the workflow around their own documents, schemas, records, and organizational processes.

This makes Darkfish useful in two ways. First, it gives users an immediate tool for processing transcripts, documents, and native-text PDFs into structured local artifacts. Second, it creates a moldable foundation that can be extended over time: new schemas, new processing functions, new artifact formats, new validation steps, and new local workflows can be added by a human developer or by an agent operating inside the user’s own local harness and permission model.

Darkfish is not designed as a cloud SaaS product. It is a locally operated tool that keeps document and PDF processing on the user’s machine by default, includes no telemetry, and contacts YouTube only when a transcript job is explicitly run. The goal is to help users preserve ownership and control over their source material while still making that material usable by machines, humans, or machine-human teams.

In short, Darkfish Foundry is a privacy-first artifact foundry and agent-composable local data toolbox. It helps users turn scattered source material into structured outputs today, while giving human-agent teams a practical framework they can refine, extend, and bend toward their own organizational needs tomorrow.

---

Start The App

Double-click:

Darkfish-Foundry.exe

The launcher starts the packaged desktop app from:

release\windows\darkfish-foundry.exe

Keep the release folder beside Darkfish-Foundry.exe; the launcher validates the nearby portable bundle before starting the app.

Included Runtime Files

Darkfish-Foundry.exe
release\windows\darkfish-foundry.exe
release\windows\darkfish.exe
release\windows\Darkfish Foundry_0.8.1-1_x64-setup.exe
release\windows\Darkfish Foundry_0.8.1-1_x64_en-US.msi
release\windows\sidecars\yt-dlp-x86_64-pc-windows-msvc.exe
release\windows\sidecars\yt-dlp.exe
release\windows\sidecars\pdfium\pdfium.dll
release\windows\sidecars\pdfium\licenses\
THIRD_PARTY_NOTICES.md
docs\

The app runtime uses the SemVer-compatible package version 0.8.1-1. The folder label v0.8.1.2 marks this cleaned user-facing package pass.

What Darkfish Does

• Processes video transcript URLs through bundled transcript sidecars.
• Processes local .txt, .md, .markdown, .json, .yaml, .yml, .csv, .log, and native-text .pdf files.
• Exports document artifacts as .txt, .md, .json, .yaml, and .docx where supported.
• Extracts native PDF text through PDFium.
• Reports scanned/image-only PDFs as ocr_required; OCR is not included in this build.
• Provides a local agent CLI and STDIO protocol through release\windows\darkfish.exe.

Privacy Boundary

Darkfish is local-first and has no telemetry.

• Local document and PDF processing stay on this machine.
• Video destinations are contacted only for transcript jobs.
• Original source files are read, not modified.
• Generated outputs are written to the selected output folder.

Output Folder

By default, Darkfish resolves the output folder dynamically for the active Windows profile:

C:\Users\<User>\Documents\Darkfish Foundry

Users can choose a custom output folder in Settings.

Agent CLI Quick Start

Use the packaged CLI:

.\release\windows\darkfish.exe doctor --json
.\release\windows\darkfish.exe process document "C:\path\source.txt" --format md --format json --metadata lite --json
.\release\windows\darkfish.exe pdf capabilities --json
.\release\windows\darkfish.exe agent --stdio

Agent-facing documentation lives under docs\.

Licensing

PDFium and other third-party notices are included in:

THIRD_PARTY_NOTICES.md
release\windows\THIRD_PARTY_NOTICES.md
release\windows\sidecars\pdfium\LICENSE-PDFIUM-BINARY.txt
release\windows\sidecars\pdfium\licenses\