Merge branch 'testing'

Author: frdel

.gitignore

@@ -4,6 +4,9 @@
 **/__pycache__/
 **/.conda/
 
+# ignore test files in root dir
+/*.test.py
+
 # Ignore git internal files (for bundler)
 .git/
 

.vscode/launch.json

@@ -7,6 +7,7 @@
       "request": "launch",
       "program": "./run_ui.py",
       "console": "integratedTerminal",
+      "justMyCode": false,
       "args": ["--development=true", "-Xfrozen_modules=off"]
     },
     {
@@ -15,6 +16,7 @@
       "request": "launch",
       "program": "./run_cli.py",
       "console": "integratedTerminal",
+      "justMyCode": false,
       "args": ["--development=true", "-Xfrozen_modules=off"]
     },
     {
@@ -23,6 +25,7 @@
       "request": "launch",
       "program": "${file}",
       "console": "integratedTerminal",
+      "justMyCode": false,
       "args": ["--development=true", "-Xfrozen_modules=off"]
     }
   ]

README.md

@@ -4,7 +4,8 @@
 
 # `Agent Zero`
 
-[![Join our Skool Community](https://img.shields.io/badge/Skool-Join%20our%20Community-4A90E2?style=for-the-badge&logo=skool&logoColor=white)](https://www.skool.com/agent-zero) [![Join our Discord](https://img.shields.io/badge/Discord-Join%20our%20server-5865F2?style=for-the-badge&logo=discord&logoColor=white)](https://discord.gg/B8KZKNsPpj) [![Subscribe on YouTube](https://img.shields.io/badge/YouTube-Subscribe-red?style=for-the-badge&logo=youtube&logoColor=white)](https://www.youtube.com/@AgentZeroFW) [![Connect on LinkedIn](https://img.shields.io/badge/LinkedIn-Connect-blue?style=for-the-badge&logo=linkedin&logoColor=white)](https://www.linkedin.com/in/jan-tomasek/) [![Follow on X.com](https://img.shields.io/badge/X.com-Follow-1DA1F2?style=for-the-badge&logo=x&logoColor=white)](https://x.com/JanTomasekDev)
+[![Thanks to Sponsors](https://img.shields.io/badge/GitHub%20Sponsors-Thanks%20to%20Sponsors-FF69B4?style=for-the-badge&logo=githubsponsors&logoColor=white)](https://github.com/sponsors/frdel) [![Join our Skool Community](https://img.shields.io/badge/Skool-Join%20our%20Community-4A90E2?style=for-the-badge&logo=skool&logoColor=white)](https://www.skool.com/agent-zero) [![Join our Discord](https://img.shields.io/badge/Discord-Join%20our%20server-5865F2?style=for-the-badge&logo=discord&logoColor=white)](https://discord.gg/B8KZKNsPpj) [![Subscribe on YouTube](https://img.shields.io/badge/YouTube-Subscribe-red?style=for-the-badge&logo=youtube&logoColor=white)](https://www.youtube.com/@AgentZeroFW) [![Connect on LinkedIn](https://img.shields.io/badge/LinkedIn-Connect-blue?style=for-the-badge&logo=linkedin&logoColor=white)](https://www.linkedin.com/in/jan-tomasek/) [![Follow on X.com](https://img.shields.io/badge/X.com-Follow-1DA1F2?style=for-the-badge&logo=x&logoColor=white)](https://x.com/JanTomasekDev)
+
 
 [Installation](./docs/installation.md) •
 [How to update](./docs/installation.md#how-to-update-agent-zero) •
@@ -13,6 +14,10 @@
 
 </div>
 
+> [!NOTE]
+> **🎉 v0.8.1 Release**: Now featuring a browser agent capable of using Chromium for web interactions! This enables Agent Zero to browse the web, gather information, and interact with web content autonomously.
+
+
 https://github.com/user-attachments/assets/c168759d-57d8-4b43-b62a-1026afcf52e6
 
 ## A personal, organic agentic framework that grows and learns with you
@@ -119,19 +124,10 @@ docker run -p 50001:80 frdel/agent-zero-run
 
 - With proper instruction, Agent Zero is capable of many things, even potentially dangerous actions concerning your computer, data, or accounts. Always run Agent Zero in an isolated environment (like Docker) and be careful what you wish for.
 
-2. **Agent Zero Is Not Pre-programmed; It Is Prompt-based.**
-
-- The whole framework contains only a minimal amount of code and does not guide the agent in any way. Everything lies in the system prompt located in the **prompts/** folder.
+2. **Agent Zero Is Prompt-based.**
 
-3. **If You Cannot Provide the Ideal Environment, Let Your Agent Know.**
+- The whole framework is guided by the **prompts/** folder. Agent guidelines, tool instructions, messages, utility AI functions, it's all there.
 
-- Agent Zero is made to be used in an isolated virtual environment (for safety) with some tools preinstalled and configured.
-
-### 📌 Known Problems
-
-1. The system prompt may need improvements; contributions are welcome!
-2. The agent may inadvertently alter its operating environment; cleaning up the `work_dir` often fixes this.
-3. Agents might loop in multi-agentic interactions, leading to unexpected behaviors.
 
 ## 📚 Read the Documentation
 
@@ -147,15 +143,19 @@ docker run -p 50001:80 frdel/agent-zero-run
 
 ### Coming soon
 
-- **User Interaction Refinements**
-- **Browser Use and RAG Tools**
+- **Knowledge and RAG Tools**
+- **Planning and Scheduling**
 
 > [!IMPORTANT]
 >
 >**Changes to frdel/agent-zero Docker image since v0.7:**
 >
 > The new Docker image `frdel/agent-zero-run` provides the new unified environment.
 
+### v0.8.1
+- **Browser Agent**
+- **UX Improvements**
+
 ### v0.8
 
 - **Docker Runtime**

agent.py

@@ -3,7 +3,7 @@
 from dataclasses import dataclass, field
 import time, importlib, inspect, os, json
 import token
-from typing import Any, Awaitable, Optional, Dict, TypedDict
+from typing import Any, Awaitable, Coroutine, Optional, Dict, TypedDict
 import uuid
 import models
 
@@ -44,7 +44,7 @@ def __init__(
         self.agent0 = agent0 or Agent(0, self.config, self)
         self.paused = paused
         self.streaming_agent = streaming_agent
-        self.process: DeferredTask | None = None
+        self.task: DeferredTask | None = None
         AgentContext._counter += 1
         self.no = AgentContext._counter
 
@@ -66,13 +66,13 @@ def first():
     @staticmethod
     def remove(id: str):
         context = AgentContext._contexts.pop(id, None)
-        if context and context.process:
-            context.process.kill()
+        if context and context.task:
+            context.task.kill()
         return context
 
     def kill_process(self):
-        if self.process:
-            self.process.kill()
+        if self.task:
+            self.task.kill()
 
     def reset(self):
         self.kill_process()
@@ -89,8 +89,8 @@ def nudge(self):
         else:
             current_agent = self.agent0
 
-        self.process = DeferredTask(current_agent.monologue)
-        return self.process
+        self.task =self.run_task(current_agent.monologue)
+        return self.task
 
     def communicate(self, msg: "UserMessage", broadcast_level: int = 1):
         self.paused = False  # unpause if paused
@@ -100,7 +100,7 @@ def communicate(self, msg: "UserMessage", broadcast_level: int = 1):
         else:
             current_agent = self.agent0
 
-        if self.process and self.process.is_alive():
+        if self.task and self.task.is_alive():
             # set intervention messages to agent(s):
             intervention_agent = current_agent
             while intervention_agent and broadcast_level != 0:
@@ -110,11 +110,19 @@ def communicate(self, msg: "UserMessage", broadcast_level: int = 1):
                     Agent.DATA_NAME_SUPERIOR, None
                 )
         else:
+            self.task = self.run_task(self._process_chain, current_agent, msg)
 
-            # self.process = DeferredTask(current_agent.monologue, msg)
-            self.process = DeferredTask(self._process_chain, current_agent, msg)
+        return self.task
 
-        return self.process
+    def run_task(
+        self, func: Callable[..., Coroutine[Any, Any, Any]], *args: Any, **kwargs: Any
+    ):
+        if not self.task:
+            self.task = DeferredTask(
+                thread_name=self.__class__.__name__,
+            )
+        self.task.start_task(func, *args, **kwargs)
+        return self.task
 
     # this wrapper ensures that superior agents are called back if the chat was loaded from file and original callstack is gone
     async def _process_chain(self, agent: "Agent", msg: "UserMessage|str", user=True):
@@ -139,18 +147,20 @@ async def _process_chain(self, agent: "Agent", msg: "UserMessage|str", user=True
 class ModelConfig:
     provider: models.ModelProvider
     name: str
-    ctx_length: int
-    limit_requests: int
-    limit_input: int
-    limit_output: int
-    kwargs: dict
+    ctx_length: int = 0
+    limit_requests: int = 0
+    limit_input: int = 0
+    limit_output: int = 0
+    vision: bool = False
+    kwargs: dict = field(default_factory=dict)
 
 
 @dataclass
 class AgentConfig:
     chat_model: ModelConfig
     utility_model: ModelConfig
     embeddings_model: ModelConfig
+    browser_model: ModelConfig
     prompts_subdir: str = ""
     memory_subdir: str = ""
     knowledge_subdirs: list[str] = field(default_factory=lambda: ["default", "custom"])
@@ -480,6 +490,30 @@ def concat_messages(
     ):  # TODO add param for message range, topic, history
         return self.history.output_text(human_label="user", ai_label="assistant")
 
+    def get_chat_model(self):
+        return models.get_model(
+            models.ModelType.CHAT,
+            self.config.chat_model.provider,
+            self.config.chat_model.name,
+            **self.config.chat_model.kwargs,
+        )
+
+    def get_utility_model(self):
+        return models.get_model(
+            models.ModelType.CHAT,
+            self.config.utility_model.provider,
+            self.config.utility_model.name,
+            **self.config.utility_model.kwargs,
+        )
+
+    def get_embedding_model(self):
+        return models.get_model(
+            models.ModelType.EMBEDDING,
+            self.config.embeddings_model.provider,
+            self.config.embeddings_model.name,
+            **self.config.embeddings_model.kwargs,
+        )
+
     async def call_utility_model(
         self,
         system: str,
@@ -494,12 +528,7 @@ async def call_utility_model(
         response = ""
 
         # model class
-        model = models.get_model(
-            models.ModelType.CHAT,
-            self.config.utility_model.provider,
-            self.config.utility_model.name,
-            **self.config.utility_model.kwargs,
-        )
+        model = self.get_utility_model()
 
         # rate limiter
         limiter = await self.rate_limiter(
@@ -526,12 +555,7 @@ async def call_chat_model(
         response = ""
 
         # model class
-        model = models.get_model(
-            models.ModelType.CHAT,
-            self.config.chat_model.provider,
-            self.config.chat_model.name,
-            **self.config.chat_model.kwargs,
-        )
+        model = self.get_chat_model()
 
         # rate limiter
         limiter = await self.rate_limiter(self.config.chat_model, prompt.format())
@@ -594,6 +618,10 @@ async def handle_intervention(self, progress: str = ""):
             await self.hist_add_user_message(msg, intervention=True)
             raise InterventionException(msg)
 
+    async def wait_if_paused(self):
+        while self.context.paused:
+            await asyncio.sleep(0.1)
+
     async def process_tools(self, msg: str):
         # search for tool usage requests in agent message
         tool_request = extract_tools.json_parse_dirty(msg)

docker/run/build.txt

@@ -15,4 +15,4 @@ docker buildx build --build-arg BRANCH=development -t frdel/agent-zero-run:devel
 docker buildx build --build-arg BRANCH=testing -t frdel/agent-zero-run:testing --platform linux/amd64,linux/arm64 --push --build-arg CACHE_DATE=$(date +%Y-%m-%d:%H:%M:%S) .
 
 # main
-docker buildx build --build-arg BRANCH=testing -t frdel/agent-zero-run:testing --platform linux/amd64,linux/arm64 --push --no-cache .
+docker buildx build --build-arg BRANCH=main -t frdel/agent-zero-run:latest --platform linux/amd64,linux/arm64 --push --build-arg CACHE_DATE=$(date +%Y-%m-%d:%H:%M:%S) .

docker/run/fs/exe/initialize.sh

@@ -21,7 +21,7 @@ apt-get update > /dev/null 2>&1 &
 /usr/sbin/sshd -D &
 
 # Start searxng server in background
-su - searxng -c "bash /exe/run_searxng.sh" &
+su - searxng -c "bash /exe/run_searxng.sh \"$@\"" &
 
 # Start A0 and restart on exit
 bash /exe/run_A0.sh "$@"

docker/run/fs/exe/run_A0.sh

@@ -1,33 +1,13 @@
 #!/bin/bash
 
-# Paths
-SOURCE_DIR="/git/agent-zero"
-TARGET_DIR="/a0"
-
-
-function setup_venv() {
-    . "/ins/setup_venv.sh" "$@"
-}
-
-function clone_repo() {
-    # Copy repository files if run_ui.py is missing in /a0 (if the volume is mounted)
-    if [ ! -f "$TARGET_DIR/run_ui.py" ]; then
-        echo "Copying files from $SOURCE_DIR to $TARGET_DIR..."
-        cp -rn --no-preserve=ownership,mode "$SOURCE_DIR/." "$TARGET_DIR"
-    fi
-}
-
-# setup and preload A0
-setup_venv
-clone_repo
-python /a0/prepare.py --dockerized=true
-python /a0/preload.py --dockerized=true
-
 # Loop to restart the Python script when it finishes
 while true; do
 
-    setup_venv
-    clone_repo
+    . "/ins/setup_venv.sh" "$@"
+    . "/ins/copy_A0.sh" "$@"
+
+    python /a0/prepare.py --dockerized=true
+    python /a0/preload.py --dockerized=true
 
     echo "Starting A0..."
     python /a0/run_ui.py \

docker/run/fs/ins/copy_A0.sh

@@ -0,0 +1,11 @@
+#!/bin/bash
+
+# Paths
+SOURCE_DIR="/git/agent-zero"
+TARGET_DIR="/a0"
+
+# Copy repository files if run_ui.py is missing in /a0 (if the volume is mounted)
+if [ ! -f "$TARGET_DIR/run_ui.py" ]; then
+    echo "Copying files from $SOURCE_DIR to $TARGET_DIR..."
+    cp -rn --no-preserve=ownership,mode "$SOURCE_DIR/." "$TARGET_DIR"
+fi

docker/run/fs/ins/install_A0.sh

@@ -13,5 +13,14 @@ git clone -b "$BRANCH" "https://github.com/frdel/agent-zero" "/git/agent-zero"
 # setup python environment
 . "/ins/setup_venv.sh" "$@"
 
+# Ensure the virtual environment and pip setup
+pip install --upgrade pip ipython requests
+
+# Install some packages in specific variants
+pip install torch --index-url https://download.pytorch.org/whl/cpu
+
+# Install remaining A0 python packages
+pip install -r /git/agent-zero/requirements.txt
+
 # Preload A0
 python /git/agent-zero/preload.py --dockerized=true

docker/run/fs/ins/install_additional.sh

@@ -1,4 +1,7 @@
 #!/bin/bash
 
 # searxng
-bash /ins/install_searxng.sh "$@"
+bash /ins/install_searxng.sh "$@"
+
+# playwright
+bash /ins/install_playwright.sh "$@"