agent-cli 0.2.0__tar.gz → 0.3.0__tar.gz

{agent_cli-0.2.0 → agent_cli-0.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agent-cli
-Version: 0.2.0
+Version: 0.3.0
 Summary: A suite of AI-powered command-line tools for text correction, audio transcription, and voice assistance.
 Author-email: Bas Nijholt <bas@nijho.lt>
 Project-URL: Homepage, https://github.com/basnijholt/agent-cli
@@ -196,13 +196,23 @@ Transcribes whatever you say into text using Wyoming ASR (Automatic Speech Recog
 │                                                   [default: 192.168.1.143]   │
 │ --asr-server-port                        INTEGER  Wyoming ASR server port.   │
 │                                                   [default: 10300]           │
-│ --clipboard            --no-clipboard             Copy transcript to         │
-│                                                   clipboard.                 │
-│                                                   [default: clipboard]       │
+│ --model            -m                    TEXT     The Ollama model to use.   │
+│                                                   Default is devstral:24b.   │
+│                                                   [default: devstral:24b]    │
+│ --ollama-host                            TEXT     The Ollama server host.    │
+│                                                   Default is                 │
+│                                                   http://localhost:11434.    │
+│                                                   [default:                  │
+│                                                   http://localhost:11434]    │
+│ --llm                  --no-llm                   Use an LLM to process the  │
+│                                                   transcript.                │
+│                                                   [default: no-llm]          │
 │ --stop                                            Stop any running           │
 │                                                   background process.        │
 │ --status                                          Check if a background      │
 │                                                   process is running.        │
+│ --clipboard            --no-clipboard             Copy result to clipboard.  │
+│                                                   [default: clipboard]       │
 │ --log-level                              TEXT     Set logging level.         │
 │                                                   [default: WARNING]         │
 │ --log-file                               TEXT     Path to a file to write    │
@@ -268,35 +278,48 @@ The process management features make it perfect for hotkey toggles. Use `--statu
  voice-assistant --stop                                                         
                                                                                 
 ╭─ Options ────────────────────────────────────────────────────────────────────╮
-│ --device-index             INTEGER  Index of the PyAudio input device to     │
-│                                     use.                                     │
-│                                     [default: None]                          │
-│ --device-name              TEXT     Device name keywords for partial         │
-│                                     matching. Supports comma-separated list  │
-│                                     where each term can partially match      │
-│                                     device names (case-insensitive). First   │
-│                                     matching device is selected.             │
-│                                     [default: None]                          │
-│ --list-devices                      List available audio input devices and   │
-│                                     exit.                                    │
-│ --asr-server-ip            TEXT     Wyoming ASR server IP address.           │
-│                                     [default: 192.168.1.143]                 │
-│ --asr-server-port          INTEGER  Wyoming ASR server port.                 │
-│                                     [default: 10300]                         │
-│ --model            -m      TEXT     The Ollama model to use. Default is      │
-│                                     devstral:24b.                            │
-│                                     [default: devstral:24b]                  │
-│ --ollama-host              TEXT     The Ollama server host. Default is       │
-│                                     http://localhost:11434.                  │
-│                                     [default: http://localhost:11434]        │
-│ --stop                              Stop any running background process.     │
-│ --status                            Check if a background process is         │
-│                                     running.                                 │
-│ --log-level                TEXT     Set logging level. [default: WARNING]    │
-│ --log-file                 TEXT     Path to a file to write logs to.         │
-│                                     [default: None]                          │
-│ --quiet            -q               Suppress console output from rich.       │
-│ --help                              Show this message and exit.              │
+│ --device-index                           INTEGER  Index of the PyAudio input │
+│                                                   device to use.             │
+│                                                   [default: None]            │
+│ --device-name                            TEXT     Device name keywords for   │
+│                                                   partial matching. Supports │
+│                                                   comma-separated list where │
+│                                                   each term can partially    │
+│                                                   match device names         │
+│                                                   (case-insensitive). First  │
+│                                                   matching device is         │
+│                                                   selected.                  │
+│                                                   [default: None]            │
+│ --list-devices                                    List available audio input │
+│                                                   devices and exit.          │
+│ --asr-server-ip                          TEXT     Wyoming ASR server IP      │
+│                                                   address.                   │
+│                                                   [default: 192.168.1.143]   │
+│ --asr-server-port                        INTEGER  Wyoming ASR server port.   │
+│                                                   [default: 10300]           │
+│ --model            -m                    TEXT     The Ollama model to use.   │
+│                                                   Default is devstral:24b.   │
+│                                                   [default: devstral:24b]    │
+│ --ollama-host                            TEXT     The Ollama server host.    │
+│                                                   Default is                 │
+│                                                   http://localhost:11434.    │
+│                                                   [default:                  │
+│                                                   http://localhost:11434]    │
+│ --stop                                            Stop any running           │
+│                                                   background process.        │
+│ --status                                          Check if a background      │
+│                                                   process is running.        │
+│ --clipboard            --no-clipboard             Copy result to clipboard.  │
+│                                                   [default: clipboard]       │
+│ --log-level                              TEXT     Set logging level.         │
+│                                                   [default: WARNING]         │
+│ --log-file                               TEXT     Path to a file to write    │
+│                                                   logs to.                   │
+│                                                   [default: None]            │
+│ --quiet            -q                             Suppress console output    │
+│                                                   from rich.                 │
+│ --help                                            Show this message and      │
+│                                                   exit.                      │
 ╰──────────────────────────────────────────────────────────────────────────────╯
 
 ```

{agent_cli-0.2.0 → agent_cli-0.3.0}/README.md

@@ -166,13 +166,23 @@ Transcribes whatever you say into text using Wyoming ASR (Automatic Speech Recog
 │                                                   [default: 192.168.1.143]   │
 │ --asr-server-port                        INTEGER  Wyoming ASR server port.   │
 │                                                   [default: 10300]           │
-│ --clipboard            --no-clipboard             Copy transcript to         │
-│                                                   clipboard.                 │
-│                                                   [default: clipboard]       │
+│ --model            -m                    TEXT     The Ollama model to use.   │
+│                                                   Default is devstral:24b.   │
+│                                                   [default: devstral:24b]    │
+│ --ollama-host                            TEXT     The Ollama server host.    │
+│                                                   Default is                 │
+│                                                   http://localhost:11434.    │
+│                                                   [default:                  │
+│                                                   http://localhost:11434]    │
+│ --llm                  --no-llm                   Use an LLM to process the  │
+│                                                   transcript.                │
+│                                                   [default: no-llm]          │
 │ --stop                                            Stop any running           │
 │                                                   background process.        │
 │ --status                                          Check if a background      │
 │                                                   process is running.        │
+│ --clipboard            --no-clipboard             Copy result to clipboard.  │
+│                                                   [default: clipboard]       │
 │ --log-level                              TEXT     Set logging level.         │
 │                                                   [default: WARNING]         │
 │ --log-file                               TEXT     Path to a file to write    │
@@ -238,35 +248,48 @@ The process management features make it perfect for hotkey toggles. Use `--statu
  voice-assistant --stop                                                         
                                                                                 
 ╭─ Options ────────────────────────────────────────────────────────────────────╮
-│ --device-index             INTEGER  Index of the PyAudio input device to     │
-│                                     use.                                     │
-│                                     [default: None]                          │
-│ --device-name              TEXT     Device name keywords for partial         │
-│                                     matching. Supports comma-separated list  │
-│                                     where each term can partially match      │
-│                                     device names (case-insensitive). First   │
-│                                     matching device is selected.             │
-│                                     [default: None]                          │
-│ --list-devices                      List available audio input devices and   │
-│                                     exit.                                    │
-│ --asr-server-ip            TEXT     Wyoming ASR server IP address.           │
-│                                     [default: 192.168.1.143]                 │
-│ --asr-server-port          INTEGER  Wyoming ASR server port.                 │
-│                                     [default: 10300]                         │
-│ --model            -m      TEXT     The Ollama model to use. Default is      │
-│                                     devstral:24b.                            │
-│                                     [default: devstral:24b]                  │
-│ --ollama-host              TEXT     The Ollama server host. Default is       │
-│                                     http://localhost:11434.                  │
-│                                     [default: http://localhost:11434]        │
-│ --stop                              Stop any running background process.     │
-│ --status                            Check if a background process is         │
-│                                     running.                                 │
-│ --log-level                TEXT     Set logging level. [default: WARNING]    │
-│ --log-file                 TEXT     Path to a file to write logs to.         │
-│                                     [default: None]                          │
-│ --quiet            -q               Suppress console output from rich.       │
-│ --help                              Show this message and exit.              │
+│ --device-index                           INTEGER  Index of the PyAudio input │
+│                                                   device to use.             │
+│                                                   [default: None]            │
+│ --device-name                            TEXT     Device name keywords for   │
+│                                                   partial matching. Supports │
+│                                                   comma-separated list where │
+│                                                   each term can partially    │
+│                                                   match device names         │
+│                                                   (case-insensitive). First  │
+│                                                   matching device is         │
+│                                                   selected.                  │
+│                                                   [default: None]            │
+│ --list-devices                                    List available audio input │
+│                                                   devices and exit.          │
+│ --asr-server-ip                          TEXT     Wyoming ASR server IP      │
+│                                                   address.                   │
+│                                                   [default: 192.168.1.143]   │
+│ --asr-server-port                        INTEGER  Wyoming ASR server port.   │
+│                                                   [default: 10300]           │
+│ --model            -m                    TEXT     The Ollama model to use.   │
+│                                                   Default is devstral:24b.   │
+│                                                   [default: devstral:24b]    │
+│ --ollama-host                            TEXT     The Ollama server host.    │
+│                                                   Default is                 │
+│                                                   http://localhost:11434.    │
+│                                                   [default:                  │
+│                                                   http://localhost:11434]    │
+│ --stop                                            Stop any running           │
+│                                                   background process.        │
+│ --status                                          Check if a background      │
+│                                                   process is running.        │
+│ --clipboard            --no-clipboard             Copy result to clipboard.  │
+│                                                   [default: clipboard]       │
+│ --log-level                              TEXT     Set logging level.         │
+│                                                   [default: WARNING]         │
+│ --log-file                               TEXT     Path to a file to write    │
+│                                                   logs to.                   │
+│                                                   [default: None]            │
+│ --quiet            -q                             Suppress console output    │
+│                                                   from rich.                 │
+│ --help                                            Show this message and      │
+│                                                   exit.                      │
 ╰──────────────────────────────────────────────────────────────────────────────╯
 
 ```

{agent_cli-0.2.0 → agent_cli-0.3.0}/agent_cli/agents/_cli_options.py

@@ -18,6 +18,11 @@ OLLAMA_HOST: str = typer.Option(
     "--ollama-host",
     help=f"The Ollama server host. Default is {config.OLLAMA_HOST}.",
 )
+LLM: bool = typer.Option(
+    False,  # noqa: FBT003
+    "--llm/--no-llm",
+    help="Use an LLM to process the transcript.",
+)
 
 
 # --- ASR (Audio) Options ---
@@ -47,11 +52,6 @@ ASR_SERVER_PORT: int = typer.Option(
     "--asr-server-port",
     help="Wyoming ASR server port.",
 )
-CLIPBOARD: bool = typer.Option(
-    True,  # noqa: FBT003
-    "--clipboard/--no-clipboard",
-    help="Copy transcript to clipboard.",
-)
 
 
 # --- Process Management Options ---
@@ -68,6 +68,11 @@ STATUS: bool = typer.Option(
 
 
 # --- General Options ---
+CLIPBOARD: bool = typer.Option(
+    True,  # noqa: FBT003
+    "--clipboard/--no-clipboard",
+    help="Copy result to clipboard.",
+)
 LOG_LEVEL: str = typer.Option(
     "WARNING",
     "--log-level",

{agent_cli-0.2.0 → agent_cli-0.3.0}/agent_cli/agents/autocorrect.py

@@ -32,7 +32,7 @@ from rich.status import Status
 
 import agent_cli.agents._cli_options as opts
 from agent_cli.cli import app, setup_logging
-from agent_cli.ollama_client import build_agent
+from agent_cli.llm import build_agent
 from agent_cli.utils import get_clipboard_text
 
 # --- Configuration ---

{agent_cli-0.2.0 → agent_cli-0.3.0}/agent_cli/agents/transcribe.py

@@ -9,23 +9,63 @@ from contextlib import AbstractContextManager, nullcontext, suppress
 import pyperclip
 from rich.console import Console
 from rich.live import Live
+from rich.panel import Panel
 from rich.text import Text
 
 import agent_cli.agents._cli_options as opts
 from agent_cli import asr, process_manager
 from agent_cli.cli import app, setup_logging
+from agent_cli.llm import process_and_update_clipboard
 from agent_cli.utils import _print, print_device_index, signal_handling_context
 
+SYSTEM_PROMPT = """
+You are an AI transcription cleanup assistant. Your purpose is to improve and refine raw speech-to-text transcriptions by correcting errors, adding proper punctuation, and enhancing readability while preserving the original meaning and intent.
+
+Your tasks include:
+- Correcting obvious speech recognition errors and mishearing
+- Adding appropriate punctuation (periods, commas, question marks, etc.)
+- Fixing capitalization where needed
+- Removing filler words, false starts, and repeated words when they clearly weren't intentional
+- Improving sentence structure and flow while maintaining the speaker's voice and meaning
+- Formatting the text for better readability
+
+Important rules:
+- Do not change the core meaning or content of the transcription
+- Do not add information that wasn't spoken
+- Do not remove content unless it's clearly an error or filler
+- Return ONLY the cleaned-up text without any explanations or commentary
+- Do not wrap your output in markdown or code blocks
+"""
+
+AGENT_INSTRUCTIONS = """
+You will be given a block of raw transcribed text enclosed in <original-text> tags, and a cleanup instruction enclosed in <instruction> tags.
+
+Your job is to process the transcribed text according to the instruction, which will typically involve:
+- Correcting speech recognition errors
+- Adding proper punctuation and capitalization
+- Removing obvious filler words and false starts
+- Improving readability while preserving meaning
+
+Return ONLY the cleaned-up text with no additional formatting or commentary.
+"""
+
+INSTRUCTION = """
+Please clean up this transcribed text by correcting any speech recognition errors, adding appropriate punctuation and capitalization, removing obvious filler words or false starts, and improving overall readability while preserving the original meaning and intent of the speaker.
+"""
+
 
 async def async_main(
+    *,
     device_index: int | None,
     device_name: str | None,
     asr_server_ip: str,
     asr_server_port: int,
-    *,
     clipboard: bool,
     quiet: bool,
     list_devices: bool,
+    model: str,
+    ollama_host: str,
+    llm: bool,
 ) -> None:
     """Async entry point, consuming parsed args."""
     logger = logging.getLogger()
@@ -54,14 +94,31 @@ async def async_main(
                 listening_message="Listening...",
             )
 
-            if transcript and clipboard:
-                pyperclip.copy(transcript)
-                logger.info("Copied transcript to clipboard.")
-                _print(console, "[italic green]Copied to clipboard.[/italic green]")
-            elif not transcript:
-                logger.info("Transcript empty.")
-            else:
-                logger.info("Clipboard copy disabled.")
+        if llm and model and ollama_host and transcript:
+            _print(console, Panel(transcript, title="[cyan]Raw Transcript 📝[/cyan]"))
+            await process_and_update_clipboard(
+                system_prompt=SYSTEM_PROMPT,
+                agent_instructions=AGENT_INSTRUCTIONS,
+                model=model,
+                ollama_host=ollama_host,
+                logger=logger,
+                console=console,
+                original_text=transcript,
+                instruction=INSTRUCTION,
+                clipboard=clipboard,
+            )
+            return
+
+        if transcript and clipboard:
+            pyperclip.copy(transcript)
+            logger.info("Copied transcript to clipboard.")
+            _print(console, "[italic blue]Copied to clipboard.[/italic blue]")
+        elif not transcript:
+            logger.info("Transcript empty.")
+        else:
+            logger.info("Clipboard copy disabled.")
+        if not quiet:
+            _print(console, f"[italic green]Transcript: {transcript}[/italic green]")
 
 
 def _maybe_live(console: Console | None) -> AbstractContextManager[Live | None]:
@@ -79,12 +136,19 @@ def transcribe(
     *,
     device_index: int | None = opts.DEVICE_INDEX,
     device_name: str | None = opts.DEVICE_NAME,
+    # ASR
     list_devices: bool = opts.LIST_DEVICES,
     asr_server_ip: str = opts.ASR_SERVER_IP,
     asr_server_port: int = opts.ASR_SERVER_PORT,
-    clipboard: bool = opts.CLIPBOARD,
+    # LLM
+    model: str = opts.MODEL,
+    ollama_host: str = opts.OLLAMA_HOST,
+    llm: bool = opts.LLM,
+    # Process control
     stop: bool = opts.STOP,
     status: bool = opts.STATUS,
+    # General
+    clipboard: bool = opts.CLIPBOARD,
     log_level: str = opts.LOG_LEVEL,
     log_file: str | None = opts.LOG_FILE,
     quiet: bool = opts.QUIET,
@@ -127,5 +191,8 @@ def transcribe(
                 clipboard=clipboard,
                 quiet=quiet,
                 list_devices=list_devices,
+                model=model,
+                ollama_host=ollama_host,
+                llm=llm,
             ),
         )

{agent_cli-0.2.0 → agent_cli-0.3.0}/agent_cli/agents/voice_assistant.py

@@ -37,20 +37,15 @@ from __future__ import annotations
 
 import asyncio
 import logging
-import sys
-import time
-from contextlib import AbstractContextManager, nullcontext, suppress
-from typing import TYPE_CHECKING
+from contextlib import suppress
 
-import pyperclip
 from rich.console import Console
 from rich.panel import Panel
-from rich.status import Status
 
 import agent_cli.agents._cli_options as opts
 from agent_cli import asr, process_manager
 from agent_cli.cli import app, setup_logging
-from agent_cli.ollama_client import build_agent
+from agent_cli.llm import process_and_update_clipboard
 from agent_cli.utils import (
     _print,
     get_clipboard_text,
@@ -58,10 +53,6 @@ from agent_cli.utils import (
     signal_handling_context,
 )
 
-if TYPE_CHECKING:
-    from pydantic_ai import Agent
-
-
 # LLM Prompts
 SYSTEM_PROMPT = """\
 You are a versatile AI text assistant. Your purpose is to either **modify** a given text or **answer questions** about it, based on a specific instruction.
@@ -86,93 +77,6 @@ Return ONLY the resulting text (either the edit or the answer), with no extra fo
 """
 
 
-# --- LLM (Editing) Logic ---
-
-INPUT_TEMPLATE = """
-<original-text>
-{original_text}
-</original-text>
-
-<instruction>
-{instruction}
-</instruction>
-"""
-
-
-async def process_with_llm(
-    agent: Agent,
-    original_text: str,
-    instruction: str,
-) -> tuple[str, float]:
-    """Run the agent asynchronously and return corrected text and elapsed time."""
-    user_input = INPUT_TEMPLATE.format(original_text=original_text, instruction=instruction)
-    t_start = time.monotonic()
-    result = await agent.run(user_input)
-    t_end = time.monotonic()
-    return result.output, t_end - t_start
-
-
-def _maybe_status(console: Console | None, model: str) -> AbstractContextManager[Status | None]:
-    """Context manager for status display."""
-    if console:
-        return Status(
-            f"[bold yellow]🤖 Applying instruction with {model}...[/bold yellow]",
-            console=console,
-        )
-    return nullcontext()
-
-
-async def process_and_update_clipboard(
-    model: str,
-    ollama_host: str,
-    logger: logging.Logger,
-    console: Console | None,
-    original_text: str,
-    instruction: str,
-) -> None:
-    """Processes the text with the LLM, updates the clipboard, and displays the result.
-
-    In quiet mode, only the result is printed to stdout.
-    """
-    agent = build_agent(
-        model=model,
-        ollama_host=ollama_host,
-        system_prompt=SYSTEM_PROMPT,
-        instructions=AGENT_INSTRUCTIONS,
-    )
-    try:
-        with _maybe_status(console, model):
-            result_text, elapsed = await process_with_llm(agent, original_text, instruction)
-
-        pyperclip.copy(result_text)
-        logger.info("Copied result to clipboard.")
-
-        if console:
-            console.print(
-                Panel(
-                    result_text,
-                    title="[bold green]✨ Result (Copied to Clipboard)[/bold green]",
-                    border_style="green",
-                    subtitle=f"[dim]took {elapsed:.2f}s[/dim]",
-                ),
-            )
-        else:
-            # Quiet mode: print result to stdout for Keyboard Maestro to capture
-            print(result_text)
-
-    except Exception as e:
-        logger.exception("An error occurred during LLM processing.")
-        _print(
-            console,
-            f"❌ [bold red]An unexpected LLM error occurred: {e}[/bold red]",
-        )
-        _print(
-            console,
-            f"   Please check your Ollama server at [cyan]{ollama_host}[/cyan]",
-        )
-        sys.exit(1)
-
-
 # --- Main Application Logic ---
 
 
@@ -186,6 +90,7 @@ async def async_main(
     asr_server_port: int,
     model: str,
     ollama_host: str,
+    clipboard: bool,
 ) -> None:
     """Main async function, consumes parsed arguments."""
     logger = logging.getLogger()
@@ -232,16 +137,22 @@ async def async_main(
             )
 
             if not instruction or not instruction.strip():
-                _print(console, "[yellow]No instruction was transcribed. Exiting.[/yellow]")
+                _print(
+                    console,
+                    "[yellow]No instruction was transcribed. Exiting.[/yellow]",
+                )
                 return
 
             await process_and_update_clipboard(
+                system_prompt=SYSTEM_PROMPT,
+                agent_instructions=AGENT_INSTRUCTIONS,
                 model=model,
                 ollama_host=ollama_host,
                 logger=logger,
                 console=console,
                 original_text=original_text,
                 instruction=instruction,
+                clipboard=clipboard,
             )
 
 
@@ -250,13 +161,18 @@ def voice_assistant(
     device_index: int | None = opts.DEVICE_INDEX,
     device_name: str | None = opts.DEVICE_NAME,
     *,
+    # ASR
     list_devices: bool = opts.LIST_DEVICES,
     asr_server_ip: str = opts.ASR_SERVER_IP,
     asr_server_port: int = opts.ASR_SERVER_PORT,
+    # LLM
     model: str = opts.MODEL,
     ollama_host: str = opts.OLLAMA_HOST,
+    # Process control
     stop: bool = opts.STOP,
     status: bool = opts.STATUS,
+    # General
+    clipboard: bool = opts.CLIPBOARD,
     log_level: str = opts.LOG_LEVEL,
     log_file: str | None = opts.LOG_FILE,
     quiet: bool = opts.QUIET,
@@ -303,5 +219,6 @@ def voice_assistant(
                 asr_server_port=asr_server_port,
                 model=model,
                 ollama_host=ollama_host,
+                clipboard=clipboard,
             ),
         )

agent_cli-0.3.0/agent_cli/llm.py

@@ -0,0 +1,141 @@
+"""Client for interacting with Ollama."""
+
+from __future__ import annotations
+
+import sys
+import time
+from contextlib import AbstractContextManager, nullcontext
+from typing import TYPE_CHECKING
+
+import pyperclip
+from pydantic_ai import Agent
+from pydantic_ai.models.openai import OpenAIModel
+from pydantic_ai.providers.openai import OpenAIProvider
+from rich.panel import Panel
+from rich.status import Status
+
+from agent_cli.utils import _print
+
+if TYPE_CHECKING:
+    import logging
+
+    from rich.console import Console
+
+
+def build_agent(
+    model: str,
+    ollama_host: str,
+    *,
+    system_prompt: str | None = None,
+    instructions: str | None = None,
+) -> Agent:
+    """Construct and return a PydanticAI agent configured for local Ollama."""
+    ollama_provider = OpenAIProvider(base_url=f"{ollama_host}/v1")
+    ollama_model = OpenAIModel(model_name=model, provider=ollama_provider)
+    return Agent(
+        model=ollama_model,
+        system_prompt=system_prompt or (),
+        instructions=instructions,
+    )
+
+
+# --- LLM (Editing) Logic ---
+
+INPUT_TEMPLATE = """
+<original-text>
+{original_text}
+</original-text>
+
+<instruction>
+{instruction}
+</instruction>
+"""
+
+
+async def process_with_llm(
+    agent: Agent,
+    original_text: str,
+    instruction: str,
+) -> tuple[str, float]:
+    """Run the agent asynchronously and return corrected text and elapsed time."""
+    user_input = INPUT_TEMPLATE.format(
+        original_text=original_text,
+        instruction=instruction,
+    )
+    t_start = time.monotonic()
+    result = await agent.run(user_input)
+    t_end = time.monotonic()
+    return result.output, t_end - t_start
+
+
+def _maybe_status(
+    console: Console | None,
+    model: str,
+) -> AbstractContextManager[Status | None]:
+    """Context manager for status display."""
+    if console:
+        return Status(
+            f"[bold yellow]🤖 Applying instruction with {model}...[/bold yellow]",
+            console=console,
+        )
+    return nullcontext()
+
+
+async def process_and_update_clipboard(
+    system_prompt: str,
+    agent_instructions: str,
+    *,
+    model: str,
+    ollama_host: str,
+    logger: logging.Logger,
+    console: Console | None,
+    original_text: str,
+    instruction: str,
+    clipboard: bool,
+) -> None:
+    """Processes the text with the LLM, updates the clipboard, and displays the result.
+
+    In quiet mode, only the result is printed to stdout.
+    """
+    agent = build_agent(
+        model=model,
+        ollama_host=ollama_host,
+        system_prompt=system_prompt,
+        instructions=agent_instructions,
+    )
+    try:
+        with _maybe_status(console, model):
+            result_text, elapsed = await process_with_llm(
+                agent,
+                original_text,
+                instruction,
+            )
+
+        if clipboard:
+            pyperclip.copy(result_text)
+            logger.info("Copied result to clipboard.")
+
+        if console:
+            console.print(
+                Panel(
+                    result_text,
+                    title="[bold green]✨ Result (Copied to Clipboard)[/bold green]",
+                    border_style="green",
+                    subtitle=f"[dim]took {elapsed:.2f}s[/dim]",
+                ),
+            )
+        else:
+            # Quiet mode: print result to stdout for Keyboard Maestro to capture
+            print(result_text)
+
+    except Exception as e:
+        logger.exception("An error occurred during LLM processing.")
+        _print(
+            console,
+            f"❌ [bold red]An unexpected LLM error occurred: {e}[/bold red]",
+        )
+        _print(
+            console,
+            f"   Please check your Ollama server at [cyan]{ollama_host}[/cyan]",
+        )
+        sys.exit(1)

{agent_cli-0.2.0 → agent_cli-0.3.0}/agent_cli.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agent-cli
-Version: 0.2.0
+Version: 0.3.0
 Summary: A suite of AI-powered command-line tools for text correction, audio transcription, and voice assistance.
 Author-email: Bas Nijholt <bas@nijho.lt>
 Project-URL: Homepage, https://github.com/basnijholt/agent-cli
@@ -196,13 +196,23 @@ Transcribes whatever you say into text using Wyoming ASR (Automatic Speech Recog
 │                                                   [default: 192.168.1.143]   │
 │ --asr-server-port                        INTEGER  Wyoming ASR server port.   │
 │                                                   [default: 10300]           │
-│ --clipboard            --no-clipboard             Copy transcript to         │
-│                                                   clipboard.                 │
-│                                                   [default: clipboard]       │
+│ --model            -m                    TEXT     The Ollama model to use.   │
+│                                                   Default is devstral:24b.   │
+│                                                   [default: devstral:24b]    │
+│ --ollama-host                            TEXT     The Ollama server host.    │
+│                                                   Default is                 │
+│                                                   http://localhost:11434.    │
+│                                                   [default:                  │
+│                                                   http://localhost:11434]    │
+│ --llm                  --no-llm                   Use an LLM to process the  │
+│                                                   transcript.                │
+│                                                   [default: no-llm]          │
 │ --stop                                            Stop any running           │
 │                                                   background process.        │
 │ --status                                          Check if a background      │
 │                                                   process is running.        │
+│ --clipboard            --no-clipboard             Copy result to clipboard.  │
+│                                                   [default: clipboard]       │
 │ --log-level                              TEXT     Set logging level.         │
 │                                                   [default: WARNING]         │
 │ --log-file                               TEXT     Path to a file to write    │
@@ -268,35 +278,48 @@ The process management features make it perfect for hotkey toggles. Use `--statu
  voice-assistant --stop                                                         
                                                                                 
 ╭─ Options ────────────────────────────────────────────────────────────────────╮
-│ --device-index             INTEGER  Index of the PyAudio input device to     │
-│                                     use.                                     │
-│                                     [default: None]                          │
-│ --device-name              TEXT     Device name keywords for partial         │
-│                                     matching. Supports comma-separated list  │
-│                                     where each term can partially match      │
-│                                     device names (case-insensitive). First   │
-│                                     matching device is selected.             │
-│                                     [default: None]                          │
-│ --list-devices                      List available audio input devices and   │
-│                                     exit.                                    │
-│ --asr-server-ip            TEXT     Wyoming ASR server IP address.           │
-│                                     [default: 192.168.1.143]                 │
-│ --asr-server-port          INTEGER  Wyoming ASR server port.                 │
-│                                     [default: 10300]                         │
-│ --model            -m      TEXT     The Ollama model to use. Default is      │
-│                                     devstral:24b.                            │
-│                                     [default: devstral:24b]                  │
-│ --ollama-host              TEXT     The Ollama server host. Default is       │
-│                                     http://localhost:11434.                  │
-│                                     [default: http://localhost:11434]        │
-│ --stop                              Stop any running background process.     │
-│ --status                            Check if a background process is         │
-│                                     running.                                 │
-│ --log-level                TEXT     Set logging level. [default: WARNING]    │
-│ --log-file                 TEXT     Path to a file to write logs to.         │
-│                                     [default: None]                          │
-│ --quiet            -q               Suppress console output from rich.       │
-│ --help                              Show this message and exit.              │
+│ --device-index                           INTEGER  Index of the PyAudio input │
+│                                                   device to use.             │
+│                                                   [default: None]            │
+│ --device-name                            TEXT     Device name keywords for   │
+│                                                   partial matching. Supports │
+│                                                   comma-separated list where │
+│                                                   each term can partially    │
+│                                                   match device names         │
+│                                                   (case-insensitive). First  │
+│                                                   matching device is         │
+│                                                   selected.                  │
+│                                                   [default: None]            │
+│ --list-devices                                    List available audio input │
+│                                                   devices and exit.          │
+│ --asr-server-ip                          TEXT     Wyoming ASR server IP      │
+│                                                   address.                   │
+│                                                   [default: 192.168.1.143]   │
+│ --asr-server-port                        INTEGER  Wyoming ASR server port.   │
+│                                                   [default: 10300]           │
+│ --model            -m                    TEXT     The Ollama model to use.   │
+│                                                   Default is devstral:24b.   │
+│                                                   [default: devstral:24b]    │
+│ --ollama-host                            TEXT     The Ollama server host.    │
+│                                                   Default is                 │
+│                                                   http://localhost:11434.    │
+│                                                   [default:                  │
+│                                                   http://localhost:11434]    │
+│ --stop                                            Stop any running           │
+│                                                   background process.        │
+│ --status                                          Check if a background      │
+│                                                   process is running.        │
+│ --clipboard            --no-clipboard             Copy result to clipboard.  │
+│                                                   [default: clipboard]       │
+│ --log-level                              TEXT     Set logging level.         │
+│                                                   [default: WARNING]         │
+│ --log-file                               TEXT     Path to a file to write    │
+│                                                   logs to.                   │
+│                                                   [default: None]            │
+│ --quiet            -q                             Suppress console output    │
+│                                                   from rich.                 │
+│ --help                                            Show this message and      │
+│                                                   exit.                      │
 ╰──────────────────────────────────────────────────────────────────────────────╯
 
 ```

{agent_cli-0.2.0 → agent_cli-0.3.0}/agent_cli.egg-info/SOURCES.txt

@@ -5,7 +5,7 @@ agent_cli/__init__.py
 agent_cli/asr.py
 agent_cli/cli.py
 agent_cli/config.py
-agent_cli/ollama_client.py
+agent_cli/llm.py
 agent_cli/process_manager.py
 agent_cli/py.typed
 agent_cli/utils.py
@@ -20,6 +20,6 @@ agent_cli/agents/_cli_options.py
 agent_cli/agents/autocorrect.py
 agent_cli/agents/transcribe.py
 agent_cli/agents/voice_assistant.py
-tests/test_ollama_client.py
+tests/test_llm.py
 tests/test_process_manager.py
 tests/test_utils.py

agent_cli-0.2.0/tests/test_ollama_client.py → agent_cli-0.3.0/tests/test_llm.py

@@ -4,7 +4,7 @@ from __future__ import annotations
 
 from typing import TYPE_CHECKING
 
-from agent_cli.ollama_client import build_agent
+from agent_cli.llm import build_agent
 
 if TYPE_CHECKING:
     import pytest

agent_cli-0.2.0/agent_cli/ollama_client.py

@@ -1,24 +0,0 @@
-"""Client for interacting with Ollama."""
-
-from __future__ import annotations
-
-from pydantic_ai import Agent
-from pydantic_ai.models.openai import OpenAIModel
-from pydantic_ai.providers.openai import OpenAIProvider
-
-
-def build_agent(
-    model: str,
-    ollama_host: str,
-    *,
-    system_prompt: str | None = None,
-    instructions: str | None = None,
-) -> Agent:
-    """Construct and return a PydanticAI agent configured for local Ollama."""
-    ollama_provider = OpenAIProvider(base_url=f"{ollama_host}/v1")
-    ollama_model = OpenAIModel(model_name=model, provider=ollama_provider)
-    return Agent(
-        model=ollama_model,
-        system_prompt=system_prompt or (),
-        instructions=instructions,
-    )