Skip to main content

A sleek (macOS) system tray application providing instant access to LLMs

Project description

AbstractAssistant

AbstractAssistant is a gateway-native desktop assistant for the AbstractFramework ecosystem.

It ships as a macOS-first tray app with a compact top-right palette and a small CLI. The desktop client stays thin: AbstractGateway owns workflow discovery, durable runs, multimodal capability defaults, provider connections, and media routing. The assistant keeps only local UX state such as sessions, window placement, downloads, and optional hotkey preferences.

High-level flow:

Tray / Palette / CLI -> AbstractGateway -> AbstractRuntime -> AbstractCore -> Providers

What You Get

  • A tray-first desktop assistant with a compact top-right query palette.
  • One published gateway assistant workflow, abstractassistant-orchestrator, as the runtime path for tray and CLI turns.
  • Gateway-backed multimodal defaults for text, voice, image, video, sound, and music routes.
  • Local microphone capture and local playback while STT/TTS execution stays on the gateway.
  • Workflow-routed image, video, sound, and music generation through gateway defaults.
  • Durable tool approvals through gateway waits instead of local side channels.
  • Gateway-backed capability-default editing, including advanced route options where the selected route supports them.

Install

pip install "abstractassistant"

Practical requirements:

  • Python 3.10+
  • An AbstractGateway instance the assistant can reach
  • macOS is the primary tray target; Linux and Windows may work but are not yet packaged to the same standard

Quick Start

Start a local gateway for development:

export ABSTRACTGATEWAY_FLOWS_DIR="$PWD/abstractgateway/flows/bundles"
export ABSTRACTGATEWAY_AUTH_TOKEN="your-shared-token"
abstractgateway serve --host 127.0.0.1 --port 8080

Launch the tray assistant:

assistant

Run a single terminal turn:

assistant run --prompt "Search the web for the latest OpenAI news and summarize it with sources."

Optional connection overrides:

assistant --gateway-url http://127.0.0.1:8080 --gateway-token "$ABSTRACTGATEWAY_AUTH_TOKEN"

The tray app does not ask you to choose a workflow. It uses the published abstractassistant-orchestrator workflow from the gateway tenant catalog.

Defaults And Durability

Gateway is the source of truth for:

  • the published abstractassistant-orchestrator workflow and its catalog entrypoint
  • provider/model defaults for multimodal routes
  • media routing and execution
  • durable run history, waits, and artifacts

The desktop client stores only local state under ~/.abstractassistant/, including:

  • sessions.json: session registry and active session
  • session transcript snapshots and last run ids
  • gateway_connection.json: gateway URL plus bearer-token or session auth state
  • local downloads and tray/palette preferences

The desktop assistant stays intentionally thinner than Flow. Settings exposes gateway capability-default routes plus a bounded advanced options JSON surface. That JSON is passed through unchanged for route features such as count, seeds, lora_adapters, guidance_2, and flow_shift when the selected route supports them.

Documentation

Start with docs/README.md.

Core guides:

Architecture records:

Development

pip install -e ".[dev]"
python -m pytest tests/basic tests/integration -q

Project Links

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

abstractassistant-0.4.10.tar.gz (278.0 kB view details)

Uploaded Source

Built Distribution

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

abstractassistant-0.4.10-py3-none-any.whl (278.0 kB view details)

Uploaded Python 3

File details

Details for the file abstractassistant-0.4.10.tar.gz.

File metadata

  • Download URL: abstractassistant-0.4.10.tar.gz
  • Upload date:
  • Size: 278.0 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for abstractassistant-0.4.10.tar.gz
Algorithm Hash digest
SHA256 f05751caee5b5b1eb522bc197a99087f4b12f2d1141829748caf96a9641951b3
MD5 bf34615fa0870e5a885c1dd75f4a52c2
BLAKE2b-256 139052340c7fea7c1586f6eb52a582ebd9c773db5ca5f009857269ab3f843f0b

See more details on using hashes here.

Provenance

The following attestation bundles were made for abstractassistant-0.4.10.tar.gz:

Publisher: release.yml on lpalbou/AbstractAssistant

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file abstractassistant-0.4.10-py3-none-any.whl.

File metadata

File hashes

Hashes for abstractassistant-0.4.10-py3-none-any.whl
Algorithm Hash digest
SHA256 abdd41725bedd0fba2fe16bc4e4fc514c33fe673c77050e1e99474b8475d4615
MD5 d90d2437287191478892df11c34ae005
BLAKE2b-256 3a9aa9ae509f66195b7405f188b3a1a36a7abcab331c5581c6da399ecb291e4a

See more details on using hashes here.

Provenance

The following attestation bundles were made for abstractassistant-0.4.10-py3-none-any.whl:

Publisher: release.yml on lpalbou/AbstractAssistant

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

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