opencode-llama-cpp-launcher 0.1.3__tar.gz → 0.1.5__tar.gz

{opencode_llama_cpp_launcher-0.1.3 → opencode_llama_cpp_launcher-0.1.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: opencode-llama-cpp-launcher
-Version: 0.1.3
+Version: 0.1.5
 Summary: One command launcher for running OpenCode with a local llama.cpp model.
 License-Expression: MIT
 License-File: LICENSE
@@ -22,52 +22,58 @@ Description-Content-Type: text/markdown
 
 # OpenCode llama.cpp Launcher
 
-A one command solution for launching [OpenCode](https://opencode.ai/) with any
-local LLM that `llama-server` can serve, including models like Qwen, DeepSeek,
-and Gemma. This launcher starts `llama-server`, waits for it to become ready,
-wires the OpenAI compatible provider config into OpenCode, and cleans up when
-the local agentic coding session ends.
+Launch [OpenCode](https://opencode.ai/) with a local model served by
+[llama.cpp](https://github.com/ggml-org/llama.cpp). The launcher starts
+`llama-server`, wires OpenCode to it, and cleans up when your session ends.
+
+![OpenCode llama.cpp Launcher demo](https://raw.githubusercontent.com/ribomo/opencode-llama-cpp-launcher/main/docs/opencode-llama-demo.gif)
 
 ## Requirements
 
-- Python 3.12+
-- OpenCode
-- llama.cpp's `llama-server`
-- A local model supported by `llama-server`, for example Qwen, DeepSeek, or
-  Gemma
+- [OpenCode](https://opencode.ai/)
+- [llama.cpp](https://github.com/ggml-org/llama.cpp)'s `llama-server`
+- A local GGUF model, such as Qwen, DeepSeek, or Gemma
 
 The launcher finds `llama-server` on `PATH`, or you can set `llama_server` in
 your config.
 
+Install OpenCode using its
+[GitHub installation instructions](https://github.com/anomalyco/opencode#installation).
+Install llama.cpp using its
+[installation guide](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md).
+
 ## Install
 
-From this repository:
+For most users, install with `pipx`:
 
 ```bash
-uv sync --dev
+pipx install opencode-llama-cpp-launcher
 ```
 
-Check that the required external binaries are available:
+Or install with `pip`:
 
 ```bash
-uv run opencode-llama doctor
+python -m pip install opencode-llama-cpp-launcher
 ```
 
-## Configure
-
-Create a project-local config in the project where you want OpenCode to run:
+Check that the required external binaries are available:
 
 ```bash
-cp opencode-llama.example.yaml opencode-llama.yaml
+opencode-llama doctor
 ```
 
-Then edit `opencode-llama.yaml`:
+## Configure
+
+Create `opencode-llama.yaml` in the project where you want OpenCode to run, or
+create `~/.config/opencode-llama.yaml` for a user-wide default:
 
 ```yaml
 model: /absolute/path/to/model.gguf
-llama_server: /optional/path/to/llama-server
-port: 8080
 ctx_size: 8192
+
+# Optional
+port: 8080
+llama_server: /optional/path/to/llama-server
 ```
 
 Config lookup order:
@@ -81,24 +87,24 @@ Config lookup order:
 Run with an explicit config file:
 
 ```bash
-uv run opencode-llama --config opencode-llama.yaml
+opencode-llama --config opencode-llama.yaml
 ```
 
 Or pass the model directly:
 
 ```bash
-uv run opencode-llama --model /absolute/path/to/model.gguf
+opencode-llama --model /absolute/path/to/model.gguf
 ```
 
 Useful options:
 
 ```bash
-uv run opencode-llama --help
-uv run opencode-llama --dry-run
-uv run opencode-llama --config opencode-llama.yaml
-uv run opencode-llama --port 9001
-uv run opencode-llama --ctx-size 8192
-uv run opencode-llama --llama-server /absolute/path/to/llama-server
+opencode-llama --help
+opencode-llama --dry-run
+opencode-llama --config opencode-llama.yaml
+opencode-llama --port 9001
+opencode-llama --ctx-size 8192
+opencode-llama --llama-server /absolute/path/to/llama-server
 ```
 
 If `llama-server` fails before becoming healthy, the launcher includes a bounded
@@ -107,6 +113,12 @@ quiet.
 
 ## Development
 
+Install dependencies from this repository:
+
+```bash
+uv sync --dev
+```
+
 Run the test suite:
 
 ```bash

opencode_llama_cpp_launcher-0.1.5/README.md

@@ -0,0 +1,117 @@
+# OpenCode llama.cpp Launcher
+
+Launch [OpenCode](https://opencode.ai/) with a local model served by
+[llama.cpp](https://github.com/ggml-org/llama.cpp). The launcher starts
+`llama-server`, wires OpenCode to it, and cleans up when your session ends.
+
+![OpenCode llama.cpp Launcher demo](https://raw.githubusercontent.com/ribomo/opencode-llama-cpp-launcher/main/docs/opencode-llama-demo.gif)
+
+## Requirements
+
+- [OpenCode](https://opencode.ai/)
+- [llama.cpp](https://github.com/ggml-org/llama.cpp)'s `llama-server`
+- A local GGUF model, such as Qwen, DeepSeek, or Gemma
+
+The launcher finds `llama-server` on `PATH`, or you can set `llama_server` in
+your config.
+
+Install OpenCode using its
+[GitHub installation instructions](https://github.com/anomalyco/opencode#installation).
+Install llama.cpp using its
+[installation guide](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md).
+
+## Install
+
+For most users, install with `pipx`:
+
+```bash
+pipx install opencode-llama-cpp-launcher
+```
+
+Or install with `pip`:
+
+```bash
+python -m pip install opencode-llama-cpp-launcher
+```
+
+Check that the required external binaries are available:
+
+```bash
+opencode-llama doctor
+```
+
+## Configure
+
+Create `opencode-llama.yaml` in the project where you want OpenCode to run, or
+create `~/.config/opencode-llama.yaml` for a user-wide default:
+
+```yaml
+model: /absolute/path/to/model.gguf
+ctx_size: 8192
+
+# Optional
+port: 8080
+llama_server: /optional/path/to/llama-server
+```
+
+Config lookup order:
+
+1. The path passed with `--config`
+2. `opencode-llama.yaml` or `opencode-llama.yml` in the project directory
+3. `~/.config/opencode-llama.yaml`
+
+## Usage
+
+Run with an explicit config file:
+
+```bash
+opencode-llama --config opencode-llama.yaml
+```
+
+Or pass the model directly:
+
+```bash
+opencode-llama --model /absolute/path/to/model.gguf
+```
+
+Useful options:
+
+```bash
+opencode-llama --help
+opencode-llama --dry-run
+opencode-llama --config opencode-llama.yaml
+opencode-llama --port 9001
+opencode-llama --ctx-size 8192
+opencode-llama --llama-server /absolute/path/to/llama-server
+```
+
+If `llama-server` fails before becoming healthy, the launcher includes a bounded
+tail of the server's startup output in the error message. Successful runs stay
+quiet.
+
+## Development
+
+Install dependencies from this repository:
+
+```bash
+uv sync --dev
+```
+
+Run the test suite:
+
+```bash
+uv run pytest
+```
+
+Before publishing, check for local files:
+
+```bash
+git status --short --ignored
+```
+
+Do not commit local launcher configs, virtual environments, caches, build
+artifacts, or model paths.
+
+## License
+
+MIT

opencode_llama_cpp_launcher-0.1.5/docs/demo.tape

@@ -0,0 +1,41 @@
+Output docs/opencode-llama-demo.gif
+
+Require opencode-llama
+Require opencode
+
+Set Shell "bash"
+Set FontSize 24
+Set FontFamily "JetBrains Mono"
+Set Width 1200
+Set Height 720
+Set Padding 18
+Set Framerate 30
+Set Theme "iTerm2 Tango Dark"
+Set BorderRadius 8
+Set TypingSpeed 25ms 
+
+Hide
+Type 'DEMO_PROJECT="$PWD"; DEMO_HOME="$(mktemp -d)"; mkdir "$DEMO_HOME/this-project"; cp -a "$DEMO_PROJECT/." "$DEMO_HOME/this-project"; cd "$DEMO_HOME"'
+Enter
+Type "export PS1='> '"
+Enter
+Type "clear"
+Enter
+Show
+
+Type "cd this-project"
+Sleep 0.5s
+Enter
+Sleep 1s
+
+Type "opencode-llama"
+Sleep 0.8s
+Enter
+Sleep 5s
+
+Type "Say hello from the local model."
+Sleep 0.8s
+Enter
+Sleep 8s
+
+Ctrl+C

{opencode_llama_cpp_launcher-0.1.3 → opencode_llama_cpp_launcher-0.1.5}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "opencode-llama-cpp-launcher"
-version = "0.1.3"
+version = "0.1.5"
 description = "One command launcher for running OpenCode with a local llama.cpp model."
 readme = "README.md"
 license = "MIT"

{opencode_llama_cpp_launcher-0.1.3 → opencode_llama_cpp_launcher-0.1.5}/uv.lock

@@ -64,7 +64,7 @@ wheels = [
 
 [[package]]
 name = "opencode-llama-cpp-launcher"
-version = "0.1.3"
+version = "0.1.5"
 source = { editable = "." }
 dependencies = [
     { name = "pyyaml" },

opencode_llama_cpp_launcher-0.1.3/README.md

@@ -1,105 +0,0 @@
-# OpenCode llama.cpp Launcher
-
-A one command solution for launching [OpenCode](https://opencode.ai/) with any
-local LLM that `llama-server` can serve, including models like Qwen, DeepSeek,
-and Gemma. This launcher starts `llama-server`, waits for it to become ready,
-wires the OpenAI compatible provider config into OpenCode, and cleans up when
-the local agentic coding session ends.
-
-## Requirements
-
-- Python 3.12+
-- OpenCode
-- llama.cpp's `llama-server`
-- A local model supported by `llama-server`, for example Qwen, DeepSeek, or
-  Gemma
-
-The launcher finds `llama-server` on `PATH`, or you can set `llama_server` in
-your config.
-
-## Install
-
-From this repository:
-
-```bash
-uv sync --dev
-```
-
-Check that the required external binaries are available:
-
-```bash
-uv run opencode-llama doctor
-```
-
-## Configure
-
-Create a project-local config in the project where you want OpenCode to run:
-
-```bash
-cp opencode-llama.example.yaml opencode-llama.yaml
-```
-
-Then edit `opencode-llama.yaml`:
-
-```yaml
-model: /absolute/path/to/model.gguf
-llama_server: /optional/path/to/llama-server
-port: 8080
-ctx_size: 8192
-```
-
-Config lookup order:
-
-1. The path passed with `--config`
-2. `opencode-llama.yaml` or `opencode-llama.yml` in the project directory
-3. `~/.config/opencode-llama.yaml`
-
-## Usage
-
-Run with an explicit config file:
-
-```bash
-uv run opencode-llama --config opencode-llama.yaml
-```
-
-Or pass the model directly:
-
-```bash
-uv run opencode-llama --model /absolute/path/to/model.gguf
-```
-
-Useful options:
-
-```bash
-uv run opencode-llama --help
-uv run opencode-llama --dry-run
-uv run opencode-llama --config opencode-llama.yaml
-uv run opencode-llama --port 9001
-uv run opencode-llama --ctx-size 8192
-uv run opencode-llama --llama-server /absolute/path/to/llama-server
-```
-
-If `llama-server` fails before becoming healthy, the launcher includes a bounded
-tail of the server's startup output in the error message. Successful runs stay
-quiet.
-
-## Development
-
-Run the test suite:
-
-```bash
-uv run pytest
-```
-
-Before publishing, check for local files:
-
-```bash
-git status --short --ignored
-```
-
-Do not commit local launcher configs, virtual environments, caches, build
-artifacts, or model paths.
-
-## License
-
-MIT