> For the complete documentation index, see [llms.txt](https://kai-7.gitbook.io/axiomkit/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://kai-7.gitbook.io/axiomkit/roadmap.md).

# Roadmap

Buillt MVP 3 month

## Month 1: Core  packages & Basic Agent Intelligence

**Goal:** Establish the absolute minimum viable core framework, allowing a simple agent to run and interact with an LLM.

### **Key Deliverables:**

* **AxiomKit Core Package :**
  * Basic **Agent** class with lifecycle methods (initialize, run, shutdown).
  * **Agent Orchestrator:** Simple event loop to process inputs and trigger agent behavior.
  * **Context Manager:** In-memory context management for a single agent instance (basic conversation history).
  * **LLM Provider Abstraction:** Interface for **LLM** providers.
  * **Action/Tool System:** Basic mechanism to define and register hardcoded actions (functions the agent can call).
  * **Input/Output Adapters (Basic CLI):** A simple CLI adapter for text input and output to interact with the agent.

### **Milestones:**

* **Week 1-2:** Set up monorepo structure, core package, basic Agent class, LLMProviders interface, and a simple CLI I/O adapter.
* **Week 3-4:** Implement basic **Agent Orchestrator** logic, integrate the chosen LLM provider, and enable the agent to receive CLI input, process it with the LLM, and output a response. Implement basic hardcoded actions .

### **Success Metric:**

A developer can run a basic AxiomKit agent via CLI, provide a prompt, and see the LLM respond, potentially triggering a simple hardcoded action.

## Month 2: **Pluggable Architecture - Sei Native Plugin**

**Goal:** Implement the core plugin system, allowing external capabilities to be easily integrated, and develop a high-value, functional plugin.

**Key Deliverables:**

* **Plugin Manager:**
  * Core logic to discover, load, and register plugins from external packages.
  * Mechanism for plugins to register their **Actions** with the **Action/Tool System**.
* **Memory System Enhancement:**
  * Introduce a basic persistent memory adapter (e.g., a simple file-based JSON store or SQLite) for long-term agent memory. This allows agents to remember across sessions.
* **First Official Plugin ( twitter  or similar high-impact integration):**
  * Define the **AgentPlugin** interface that plugins must adhere to.
  * Implement **at least two core actions** for the chosen plugin (e.g., **createPost**, **readThePost , reply for twitter**).
  * Handle necessary authentication/API key management for the plugin (e.g., environment variables).
* **CLI Enhancements:**
  * **axiomkit add plugin \<plugin-name>** command to simplify adding plugins to a project.
* Sei Native Plugin (Two Core Functions):
  * **Send Native Tokens:** An action to programmatically send SEI tokens.
  * **Query Sei Chain Data:** An action to read basic on-chain data (e.g., account balance, simple contract state).

**Milestones:**

* **Week 5-6:** Develop the **Plugin Manager** within **@axiomkit/core**. Refine **Action/Tool System** to accept plugin-provided actions.
* **Week 7-8:** Develop the initial `plugin`package. Implement its core actions and ensure it registers correctly with the Plugin Manager. Integrate the persistent memory adapter. Update the CLI to support adding plugins.

**Success Metric:** A developer can install the twitter bot plugin, configure it, and the agent, using the LLM, can successfully send and/or read emails based on a prompt.

## **Month 3: Developer Experience (DX) & MVP Solidification**

**Goal:** Greatly enhance the developer experience, provide initial documentation, and solidify the MVP for public release.

### **Key Deliverables:**

* Basic UI Support:
  * Allow developers easier to view
* **AxiomKit CLI Project Generator (`@axiomkit/cli`):**
  * `axiomkit new <project-name>` command to scaffold a new AxiomKit agent project with a basic setup (core, recommended dependencies, example agent file).
* **Basic Observability & Debugging:**
  * Implement structured logging within the `Core Runtime` to track LLM calls, action executions, and key agent decisions (output to console/file). This forms the basis of the "Agent Flight Recorder."
* **Initial Documentation (GitBook):**
  * **Introduction:** Welcome, Why AxiomKit (MVP differentiators).
  * **Getting Started:** Installation, Your First Agent (CLI generated).
  * **Core Concepts:** Basic explanations of Agent, Context, Action, Memory, LLM Provider.
  * **Plugin Guide:** How to use the first official plugin.
  * **Contribution Guide (Basic):** How to set up the dev environment and run tests.
* **Testing Suite:**
  * Basic unit tests for core components (Agent Orchestrator, Plugin Manager, Action System).
  * Integration tests for the LLM Provider and the first official plugin.
* **Refinement & Hardening:**
  * Address critical bugs found during development and testing.
  * Basic error handling for common scenarios.
  * Review code for clarity and maintainability.

### **Milestones:**

* **Week 9-10:** Develop the `axiomkit new` CLI command. Implement initial logging and observability features.
* **Week 11-12:** Write comprehensive GitBook documentation for all MVP features. Develop initial test suite. Perform final bug fixes and polish for MVP release.

### **Success Metric:**&#x20;

A new developer can follow the GitBook documentation, use the CLI to generate a project, install the twitter plugin, and successfully run an agent that demonstrates the core features. The logging provides clear insights into the agent's operations.