v0.0.15

Plandex integration

# Run plandex service
harbor up plandex
# Run healthcheck against the Plandex service
harbor plandex health # should print "OK"

• Service overview
• Setup/configuration with Ollama guide in the Wiki
• harbor pdx CLI reference

Local env

.env file can now be used for any necessary permanent overrides without risking to conflict with updates from upstream.

To accompany the change:

# Brings the configuration back to current defaults
harbor config reset

Misc

• HuggingFace CLI is a service - unify config management with other services
• Poor man's harbor update

---

Full Changelog: v0.0.14...v0.0.15

v0.0.14

PAR LLAMA

TUI for Ollama.

PAR LLAMA is an amazing terminal UI for interacting with Ollama. Everything you'd expect from a modern chat interface, but in the terminal.

# Ollama should be one of the
# running services in order to be reachable by parllama
harbor up ollama

# 1. Shortcut
harbor parllama

# 2. Via harbor run - the underlying command is different
harbor run parllama

# 3. Via interactive service shell
harbor shell parllama
$ parllama

See service wiki for more details

Full Changelog: v0.0.13...v0.0.14

v0.0.13

What's Changed

TabbyAPI backend support

Wiki Docs

harbor tabbyapi model Annuvin/gemma-2-2b-it-abliterated-4.0bpw-exl2
harbor up tabbyapi

New CLI Features

harbor hf dl

Integrating awesome HuggingFaceModelDownloader CLI for easier HF/Llama.cpp cache management

# See the original help
harbor hf dl --help

# EXL2 example
#
# -s ./hf - Save the model to global HuggingFace cache (mounted to ./hf)
# -c 10   - make download go brr with 10 concurrent connections
# -m      - model specifier in user/repo format
# -b      - model revision/branch specifier (where applicable)
harbor hf dl -c 10 -m turboderp/TinyLlama-1B-exl2 -b 2.3bpw -s ./hf

# GGUF example
#
# -s ./llama.cpp - Save the model to global llama.cpp cache (mounted to ./llama.cpp)
# -c 10          - make download go brr with 10 concurrent connections
# -m             - model specifier in user/repo format
# :Q2_K          - file filter postfix - will only download files with this postfix
harbor hf dl -c 10 -m TheBloke/TinyLlama-1.1B-Chat-v1.0-GGUF:Q2_K -s ./llama.cpp

harbor hf find

To accompany the hf dl - a quick way to jump right to the HF Hub to find new models.

harbor hf find gguf
harbor hf find exl2 gemma-2
harbor hf find awq llama-3.1
harbor hf find tinyllama

Misc

• docs: update README.md by @eltociear in #3
• harbor shell - launch interactive shell in service container (shortcut from previous harbor exec + harbor cmd combinations)
• harbor build - for services that'll have their Dockerfile within Harbor repo (such as hfdownloader)

New Contributors

• @eltociear made their first contribution in #3

Full Changelog: v0.0.12...v0.0.13

v0.0.12

• Container name prefix: all containers from the toolkit are now prefixed with harbor. to avoid conflict with other locally running setups

# Can be adjusted via .env or with CLI
harbor config get container.prefix # harbor
harbor config set container.prefix friendly # friendly.ollama

• Fixing harbor ln to use CLI name without an extension

Full Changelog: v0.0.11...v0.0.12

v0.0.11

Features

Open WebUI cross-file configuration

Setup for Open WebUI was refactored to actively intersect configs from requested services. In other words - now Open WebUI won't try to reach out to LiteLLM or vLLM (or other) service endpoints unless it has been explicitly launched

OpenAI Key/URL management

harbor openai urls add https://api.openai.com/v1
harbor openai keys add <OpenAI API Key>

Default services management

harbor defaults ls
harbor defaults rm 0
harbor defaults rm ollama
harbor defaults add llamacpp

Autoopen

# Update the config
harbor config set ui.autoopen true

# Start harbor - default UI opens automatically
harbor up

Misc

• Compatibility Wiki
• env manager works with array variables
• harbor webui CLI helper
• harbor dive util
• harbor info util

---

Full Changelog: v0.0.10...v0.0.11

v0.0.10

• harbor up aphrodite - Aphrodite Engine inference backend

Full Changelog: v0.0.9...v0.0.10

v0.0.9

• harbor up vllm - vLLM backend support
• harbor up bionicgpt - BionicGPT frontend
• litellm config merger workflow
• harbor cmd - run arbitrary compose commands
• harbor run - run commands in a service main container without starting it first

Full Changelog: v0.0.8...v0.0.9

v0.0.8

• harbor up librechat - LibreChat as an additional UI service
• harbor pull to update service images
• harbor down can now pass parameters to the docker

Full Changelog: v0.0.7...v0.0.8

v0.0.7

• 📖 Wiki
• langfuse service, sample integration with litellm
• hollama frontend
• crossfile support for compose

Full Changelog: v0.0.6...v0.0.7

v0.0.6

• Approaching v0.1.0 with only few services remaining to scaffold 🎉
• harbor up tts - TTS integrated with Open WebUI
• harbor up tgi litellm - TGI as another LLM backend and LiteLLM as an OpenAI-compatible API proxy