-### **Developer Experience**
-- **Plug-and-Play Nodes**: Create reusable, atomic workflow components that can be mixed and matched
-- **Python-First**: Native Python support with Pydantic models for type-safe inputs/outputs
+
-### **Production Infrastructure**
-- **Kubernetes Native**: Deploy seamlessly on Kubernetes clusters for enterprise-grade scalability
-- **State Persistence**: Maintain workflow state across restarts and failures
-- **Interactive Dashboard**: Visual workflow management, monitoring, and debugging tools
+### [Lightweight Runtime](./exosphere/architecture.md)
-### **AI Agent Capabilities**
-- **Autonomous Execution**: Build agents that can make decisions and execute complex workflows
-- **Data Processing**: Handle large datasets with distributed processing capabilities
-- **API Integration**: Connect to external services and APIs through configurable nodes
+Execute workflows reliably with minimal overhead across distributed infrastructure using a state-based execution model.
-Whether you're building data pipelines, AI agents, or complex workflow orchestrations, Exosphere provides the infrastructure backbone to make your AI applications production-ready and scalable.
+
----
+
-**Watch the Step by Step Demo**:
+### [Inbuilt Failure Handling](./exosphere/retry-policy.md)
-
- 
-
+Built-in retry policies with exponential backoff and jitter strategies for resilient, production-grade execution.
+
----
+
-## Requirements
+### [Infinite Parallel Agents](./exosphere/fanout.md)
-Python 3.12+
+Scale to unlimited parallel agents with automatic load distribution and dynamic fanout at runtime.
-## Installation
+
-```bash
-uv add exospherehost
-```
+
-## Example
+### [Dynamic Execution Graphs](./exosphere/concepts.md)
-### Create
+Durable execution designed for agentic flows with node based control of execution.
-Create a file `main.py` with:
+
-```python
-from exospherehost import Runtime, BaseNode
-from pydantic import BaseModel
+
-class HelloWorldNode(BaseNode):
- class Inputs(BaseModel):
- name: str
+### [Native State Persistence](./exosphere/store.md)
- class Outputs(BaseModel):
- message: str
+Persist workflow state across restarts and failures with graph-level key-value storage.
- class Secrets(BaseModel):
- pass
+
- async def execute(self) -> Outputs:
- return self.Outputs(
- message=f"Hello, {self.inputs.name}!"
- )
+
-# Initialize the runtime
-Runtime(
- namespace="MyProject",
- name="HelloWorld",
- nodes=[HelloWorldNode]
-).start() # Note: This blocks the main thread
-```
+### [Observability](./exosphere/dashboard.md)
-!!! info "Interactive Environments"
- If you're using Jupyter notebooks or Python REPLs, `Runtime.start()` will block your session. See the [Getting Started guide](./getting-started.md#important-blocking-behavior) for non-blocking alternatives.
+Visual monitoring, debugging, and management of workflows with real-time execution tracking.
-### Run
+
-Run the server with:
+
-Ref: [Local Setup](./exosphere/local-setup.md)
+
-## Example graph
+### [Getting Started](./getting-started.md)
-Now modify the file `main.py` to add more complex processing:
+Get the Exosphere State Manager and Dashboard running locally for development.
-```python
-from exospherehost import Runtime, BaseNode
-from pydantic import BaseModel
-import json
+
-class DataProcessorNode(BaseNode):
- class Inputs(BaseModel):
- data: str
- operation: str
+
- class Outputs(BaseModel):
- result: str
- status: str
+### [Run your first Node](./exosphere/register-node.md)
- class Secrets(BaseModel):
- api_key: str
+Create your first node and register it with the Exosphere runtime.
- async def execute(self) -> Outputs:
- # Parse the input data
- try:
- data = json.loads(self.inputs.data)
- except json.JSONDecodeError as e:
- return self.Outputs(
- result="",
- status=f"error: invalid json (line {e.lineno}, column {e.colno})"
- )
+
- # Process based on operation
- if self.inputs.operation == "transform":
- result = {"transformed": data, "processed": True}
- else:
- result = {"original": data, "processed": False}
+
- return self.Outputs(result=json.dumps(result), status="success")
+### [Trigger Agent](./exosphere/trigger-graph.md)
+Learn how to trigger your agent workflows and manage execution flows.
-# Initialize the runtime
-Runtime(
- namespace="MyProject",
- name="DataProcessor",
- nodes=[DataProcessorNode]
-).start()
-```
+
-The runtime will automatically reload and register the updated node.
+
+
+### [Deploy & Monitor](./exosphere/dashboard.md)
+
+Deploy your agents and monitor their execution with the visual dashboard.
+
+
+
+
+