Skip to main content

Sync email threads from IMAP to Markdown, draft replies, manage collaborators

Project description

Corky logo

Corky

Alpha software. Expect breaking changes between minor versions. See VERSIONS.md for migration notes.

Full documentation: https://btakita.github.io/corky

Sync email threads from IMAP to Markdown. Draft replies with AI assistance. Share scoped threads with collaborators via git.

Corky syncs threads from any IMAP provider (Gmail, Protonmail Bridge, self-hosted) into mail/conversations/ — one file per thread, regardless of source. A thread that arrives via both Gmail and Protonmail merges into one file. Labels, accounts, and contacts are metadata, not directory structure.

Install

pip install corky        # or: pipx install corky

Or via shell installer:

curl -sSf https://raw.githubusercontent.com/btakita/corky/main/install.sh | sh

Or from source: cargo install --path .

Quick start

corky init --user you@gmail.com
# Edit mail/.corky.toml with credentials
corky sync

See the getting started guide for full setup instructions.

Key features

  • Flat conversations — one Markdown file per thread, all sources merged
  • Sandboxed sharing — label-based routing gives collaborators only the threads you choose
  • AI-native — files, CLI, and git work the same for humans and agents
  • Multi-account — Gmail, Protonmail Bridge, generic IMAP, all in one directory
  • Social posting — draft and publish to LinkedIn (and future platforms) via OAuth
  • Scheduling — schedule email and social drafts for timed publishing
  • Topics — organize conversations with shared topic context across mailboxes
  • Transcription — whisper-rs audio transcription with speaker diarization via pyannote-rs
  • Watch daemon — poll IMAP and run scheduled publishing with corky watch

Usage

corky sync                      # Incremental IMAP sync
corky unanswered                # Threads awaiting a reply
corky draft push FILE           # Save as email draft
corky mailbox add NAME --label LABEL  # Share threads
corky contact sync              # Sync contact CLAUDE.md between root and mailboxes
corky filter push               # Push Gmail filters from .corky.toml
corky filter push --dry-run     # Preview filter changes
corky filter pull               # Show current Gmail filters
corky filter auth               # Authenticate for Gmail filter API
corky linkedin draft              # Create LinkedIn draft
corky linkedin publish FILE      # Publish to LinkedIn
corky schedule run              # Publish due scheduled items
corky topics list               # Show configured topics
corky watch                     # Poll, sync, and publish scheduled
corky transcribe FILE            # Transcribe audio to text
corky transcribe FILE --diarize  # With speaker diarization
corky --help                    # All commands

Transcription & speaker diarization

Transcribe audio files with optional speaker diarization. Supports WAV, MP3, FLAC, OGG, M4A, AMR, and more.

# Basic transcription
corky transcribe call.amr -o transcript.md

# With speaker diarization (interactive labeling)
corky transcribe call.amr --diarize -o transcript.md

# With pre-assigned speaker names
corky transcribe call.amr --diarize --speakers "Alice,Bob" -o transcript.md

Diarization uses pyannote-rs (ONNX Runtime) to detect and label speakers. When run without --speakers, corky shows text excerpts per speaker and prompts you to assign names interactively. ONNX models auto-download on first use — no gated HuggingFace access required.

Feature flags: Transcription (CPU) is enabled by default. For GPU acceleration, install with --features transcribe-cuda. Diarization requires --features diarize.

This feature was designed collaboratively using agent-doc interactive document sessions.

Telegram import

Import Telegram Desktop JSON exports into corky conversations:

corky sync telegram-import ~/Downloads/telegram-export/result.json
corky sync telegram-import ~/Downloads/telegram-export/           # directory of exports
corky sync telegram-import result.json --label personal --account tg-personal

Export from Telegram Desktop: Settings > Advanced > Export Telegram data > JSON format.

Slack import

Import Slack workspace export ZIPs:

corky slack import ~/Downloads/my-workspace-export.zip
corky slack import export.zip --label work --account slack-work

Export from Slack: Workspace admin > Settings > Import/Export Data > Export.

See the command reference for details.

Development

cp .corky.toml.example mail/.corky.toml
make check    # clippy + test
make release  # build + symlink to .bin/corky

See building and conventions.

AI agent instructions

Project instructions live in AGENTS.md (symlinked as CLAUDE.md). Personal overrides go in CLAUDE.local.md / AGENTS.local.md (gitignored).

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

corky-0.16.0.tar.gz (298.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

corky-0.16.0-py3-none-manylinux_2_39_x86_64.whl (8.3 MB view details)

Uploaded Python 3manylinux: glibc 2.39+ x86-64

File details

Details for the file corky-0.16.0.tar.gz.

File metadata

  • Download URL: corky-0.16.0.tar.gz
  • Upload date:
  • Size: 298.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: maturin/1.12.4

File hashes

Hashes for corky-0.16.0.tar.gz
Algorithm Hash digest
SHA256 efb25e9a5066731c2d2883993e5289ff08187f857e88529d04a1595f6bf2c062
MD5 80f6edf801650688a6e67dd6b85e7b4f
BLAKE2b-256 d9bef2752533490d7d8667c29400f595dd26f32bc54d5f62c53ab00f8ebf6c72

See more details on using hashes here.

File details

Details for the file corky-0.16.0-py3-none-manylinux_2_39_x86_64.whl.

File metadata

File hashes

Hashes for corky-0.16.0-py3-none-manylinux_2_39_x86_64.whl
Algorithm Hash digest
SHA256 609f461e28de259e34d1537c9dccde3e4b19acf20cf3397873442a91ca4913eb
MD5 cae28c9d088463f036c81202c286cde9
BLAKE2b-256 84072df7baccd59c6698789d2692a2d41d0195f511e34621885f4488abf7e449

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page