chore: docs (#81)

* refactor: tests

* wip: chore docs

* wip: chore docs

* docs: readmes

* chore(cza): update templates

* fix: build

* fix: anthropic

* fix(anthropic): zee

* nit

* feat(zee): image handling

* fic

---------

Co-authored-by: Ganesh Swami <[email protected]>

Author: jagnani73

README.md

@@ -2,7 +2,7 @@
 
 # AI Agent SDK for TypeScript
 
-[📖 Documentation](https://cxt.build/) | 
+[📖 Documentation](https://cxt.build/) |
 [✍🏻 ZEE Use-cases](https://cxt.build/docs/use-cases/overview)
 
 <br />
@@ -15,7 +15,6 @@
 [![GitHub issues](https://img.shields.io/github/issues/covalenthq/ai-agent-sdk)](https://github.com/covalenthq/ai-agent-sdk/issues)
 [![GitHub pull requests](https://img.shields.io/github/issues-pr/covalenthq/ai-agent-sdk)](https://github.com/covalenthq/ai-agent-sdk/pulls)
 
-
 [![GitHub stars](https://img.shields.io/github/stars/covalenthq/ai-agent-sdk)](https://github.com/covalenthq/ai-agent-sdk/stargazers)
 [![GitHub forks](https://img.shields.io/github/forks/covalenthq/ai-agent-sdk)](https://github.com/covalenthq/ai-agent-sdk/network/members)
 
@@ -25,10 +24,10 @@
 
 ## Features
 
-- LLMs - a unified interface for all LLMs
-- Agents - a single model with a system prompt and a set of tools
-- Tools - extend the capabilities of agents with external tools
-- ZEE Workflows - compose agents to solve complex problems
+-   LLMs - a unified interface for all LLMs
+-   Agents - a single model with a system prompt and a set of tools
+-   Tools - extend the capabilities of agents with external tools
+-   ZEE Workflows - compose agents to solve complex problems
 
 ## Using the SDK (Quick Start)
 
@@ -45,27 +44,31 @@ const agent1 = new Agent({
     name: "Agent1",
     model: {
         provider: "OPEN_AI",
-        name: "gpt-4o-mini",
+        id: "gpt-4o-mini",
     },
     description: "A helpful AI assistant that can engage in conversation.",
+    instructions: ["Interact with the user in a friendly and helpful manner"],
 });
 ```
 
 ### 3. Modify the ZEE Workflow
 
 ```js
 const zee = new ZeeWorkflow({
-    description: "A workflow of agents that do stuff together",
-    output: "Just bunch of stuff",
-    agents: { agent1, agent2 },
+    goal: "A workflow of agents that do stuff together",
+    agents: [agent1, agent2],
+    model: {
+        provider: "OPEN_AI",
+        id: "gpt-4o-mini",
+    },
 });
 ```
 
 ### 4. Run the Zee Workflow
 
 ```js
 (async function main() {
-    const result = await ZeeWorkflow.run(zee);
+    const result = await zee.run();
     console.log(result);
 })();
 ```

docs/concepts/agents.mdx

@@ -16,36 +16,45 @@ The simplest way to create an Agent is to use initialize an Agent with a model,
 const agent = new Agent({
     name: "Reporting Agent",
     model: {
-        provider: "OPEN_AI",
-        name: "gpt-4o-mini",
+        provider: "openai",
+        id: "gpt-4o-mini",
     },
     description: "This agent is responsible for generating reports",
-    instructions: ["Generate a report on the current state of the company"],
+    instructions: [
+        "Generate a report on the current state of the company",
+        "Use the following tools to help you generate the report",
+    ],
 });
 ```
 
 ## Adding tools
 
-[Tools](/concepts/tools) are optional and can be added to an Agent to extend its capabilities. Tools are included in the calls to the language model through features like OpenAI's [function calling](https://platform.openai.com/docs/guides/function-calling) or Claude's [tool use](https://docs.anthropic.com/en/docs/build-with-claude/tool-use).
+[Tools](/concepts/tools) are optional and can be added to an Agent to extend its capabilities. Tools are included in the calls to the language model through the [tools](https://sdk.vercel.ai/docs/ai-sdk-core/tools-and-tool-calling) property.
 
-Tools are created using the `createTool` function and can be sent to the agent using the `tools` property during initialization.
+Tools are created using the `Tool` class and can be sent to the agent using the `tools` property during initialization.
 
 ```typescript
 const agent = new Agent({
     name: "Reporting Agent",
     ...
-    tools: [tool1, tool2, tool3],
+    tools: {
+        tool1,
+        tool2,
+        tool3,
+    },
 });
 ```
 
 Read more about [Tools](/concepts/tools) to learn how to create them.
 
 ## Running an Agent
 
-Agents can be run standalone or as part of a ZEE workflow. Running an agent standalone is useful for testing and debugging. The `run` method will return the final state of the agent.
+Agents can be run standalone or as part of a ZEE workflow. Running an agent standalone is useful for testing and debugging. The `generate` method will run the agent standalone.
 
 ```typescript
-const result = await agent.run();
+const result = await agent.generate({
+    messages: [userMessage("What is the current state of the company?")],
+});
 ```
 
-Running an agent as part of a ZEE workflow is useful for solving complex problems. Read more about [ZEE](/concepts/zeeworkflows) to learn how to run agents in a ZEE workflow.
+Running an agent as part of a ZEE workflow is useful for solving complex problems. Read more about [ZEE](/concepts/zee-workflows) to learn how to run agents in a ZEE workflow.

docs/concepts/llms.mdx

@@ -4,45 +4,72 @@ description: Overview of supported language models, including OpenAI, Anthropic,
 icon: "computer-classic"
 ---
 
-Each LLM is an adapter around a language model provider and the specific model version, eg: `gpt-4o-mini`. Each [Agent](./agents) can pick their own model and a [ZeeWorkflow](./zeeworkflows) can be configured to use a specific LLM as default.
+Each LLM is an adapter around a language model provider and the specific model version, eg: `gpt-4o-mini`. Each [Agent](/agents) can pick their own model and a [ZeeWorkflow](/zee-workflows) can be configured to use a specific LLM as default.
 
 ```tsx
 const llm = new LLM({
-    provider: "OPEN_AI",
-    name: "gpt-4o-mini",
-    apiKey: process.env.OPENAI_API_KEY,
-    temperature: 0.7
+    provider: "openai",
+    id: "gpt-4o-mini",
 });
 ```
 
-## List of supported LLMs
-
-### Open AI
+## List of supported LLM Model IDs
 
 <CodeGroup>
 
-```plaintext OpenAI
-"gpt-4"
-"gpt-4-turbo"
-"gpt-3.5-turbo"
+```plaintext openai
 "gpt-4o"
 "gpt-4o-mini"
-"o3-mini"
-```
-
-```plaintext DeepSeek
-"deepseek-chat"
-"deepseek-coder"
+"gpt-4o-2024-05-13"
+"gpt-4o-2024-08-06"
+"gpt-4o-2024-11-20"
+"gpt-4o-audio-preview"
+"gpt-4o-audio-preview-2024-10-01"
+"gpt-4o-audio-preview-2024-12-17"
+"gpt-4o-mini-2024-07-18"
+"gpt-4-turbo"
+"gpt-4-turbo-2024-04-09"
+"gpt-4-turbo-preview"
+"gpt-4-0125-preview"
+"gpt-4-1106-preview"
+"gpt-4"
+"gpt-4-0613"
+"gpt-3.5-turbo-0125"
+"gpt-3.5-turbo"
+"gpt-3.5-turbo-1106"
 ```
 
-```plaintext Grok
-"grok-2-latest"
-"grok-beta"
+```plaintext anthropic
+"claude-3-5-sonnet-latest"
+"claude-3-5-sonnet-20241022"
+"claude-3-5-sonnet-20240620"
+"claude-3-5-haiku-latest"
+"claude-3-5-haiku-20241022"
+"claude-3-opus-latest"
+"claude-3-opus-20240229"
+"claude-3-sonnet-20240229"
+"claude-3-haiku-20240307"
 ```
 
-```plaintext Gemini
+```plaintext google
+"gemini-2.0-flash-001"
 "gemini-1.5-flash"
+"gemini-1.5-flash-latest"
+"gemini-1.5-flash-001"
+"gemini-1.5-flash-002"
+"gemini-1.5-flash-8b"
+"gemini-1.5-flash-8b-latest"
+"gemini-1.5-flash-8b-001"
 "gemini-1.5-pro"
+"gemini-1.5-pro-latest"
+"gemini-1.5-pro-001"
+"gemini-1.5-pro-002"
+"gemini-2.0-flash-lite-preview-02-05"
+"gemini-2.0-pro-exp-02-05"
+"gemini-2.0-flash-thinking-exp-01-21"
+"gemini-2.0-flash-exp"
+"gemini-exp-1206"
+"learnlm-1.5-pro-experimental"
 ```
 
 </CodeGroup>
@@ -51,20 +78,16 @@ const llm = new LLM({
 
 <CodeGroup>
 
-```plaintext OpenAI
+```plaintext openai
 OPENAI_API_KEY
 ```
 
-```plaintext DeepSeek
-DEEPSEEK_API_KEY
-```
-
-```plaintext Grok
-GROK_API_KEY
+```plaintext anthropic
+ANTHROPIC_API_KEY
 ```
 
-```plaintext Gemini
-GEMINI_API_KEY
+```plaintext google
+GOOGLE_GENERATIVE_AI_API_KEY
 ```
 
 </CodeGroup>
@@ -77,57 +100,41 @@ LLMs can also process images along with text using image URL messages. Here's an
 
 ```typescript
 const messages = [
-    {
-        role: "user",
-        content: [
-            {
-                type: "text",
-                text: "What's in this image? Analyze the logo and suggest improvements.",
-            },
-            {
-                type: "image_url",
-                image_url: {
-                    url: "https://example.com/logo.png",
-                    detail: "auto",
-                },
-            },
-        ],
-    },
-];
-
-const schema = {
-    analysis: z.object({
-        description: z.string(),
-        colors: z.array(z.string()),
-        text_content: z.string().optional(),
-        improvements: z.string().optional(),
-    }),
-};
+    userMessage(
+        "What's in this image?"
+    ),
+    userMessage([
+        {
+            type: "image",
+            image: "https://example.com/logo.png",
+        },
+    ]),
+],
+
+const schema = z.object({
+    description: z.string(),
+    colors: z.array(z.string()),
+    text_content: z.string().optional(),
+    improvements: z.string().optional(),
+});
 
-const result = await llm.generate(messages, schema, {});
+const result = await llm.generate({
+    messages,
+    schema,
+    temperature: 0.8,
+});
 ```
 
 The LLM will analyze the image and return a structured response containing the description, colors used, and potential improvements. You can also use base64-encoded images by providing the data URL:
 
 ```typescript
 const messages = [
-    {
-        role: "user",
-        content: [
-            {
-                type: "text",
-                text: "What's in this image and what color is it?",
-            },
-            {
-                type: "image_url",
-                image_url: {
-                    url: "data:image/png;base64,..." // Your base64 image data
-                    detail: "auto",
-                },
-            },
-        ],
-    },
+    userMessage("What's in this image?"),
+    userMessage([
+        {
+            type: "image",
+            image: "data:image/png;base64,...",
+        },
+    ]),
 ];
 ```
-
-Note: Image analysis is currently supported by OpenAI models. Some providers like Gemini may have limited or no support for image analysis.

docs/concepts/tools.mdx

@@ -10,23 +10,26 @@ Tools are functions that can be used to extend the capabilities of your agents.
 
 2. **Structured Output**: Tools are functions that can be used to generate structured output. This is useful for generating JSON, CSV, or other structured data.
 
-Tools are optional and can be added to an Agent to extend its capabilities. Tools are included in the calls to the language model through features like OpenAI's [function calling](https://platform.openai.com/docs/guides/function-calling) or Claude's [tool use](https://docs.anthropic.com/en/docs/build-with-claude/tool-use).
+Tools are optional and can be added to an Agent to extend its capabilities. Tools are included in the calls to the language model through the [tools](https://sdk.vercel.ai/docs/ai-sdk-core/tools-and-tool-calling) property.
 
 ## Creating a Tool
 
-Tools are created using the `createTool` function and can be sent to the agent using the `tools` property during initialization.
+Tools are created using the `Tool` class and can be sent to the agent using the `tools` property during initialization.
 
 ```typescript
-const tool = createTool({
-    id: "get-company-report",
-    description: "This tool is used to get the current state of the company",
-    schema: z.object({}),
-    execute: async (params) => {
-        return "The current state of the company is good";
+const tool = new Tool({
+    provider: "openai",
+    name: "get weather",
+    description: "This tool is used to get the current weather in a location",
+    parameters: z.object({
+        location: z.string(),
+    }),
+    execute: async ({ location }) => {
+        return "The current weather in " + location + " is sunny";
     },
 });
 ```
 
-## Available Tools
+## Default Tools
 
 - [GoldRush Onchain Data](/tools/goldrush-onchain-data) - Tools for interacting with blockchain data