> ## Documentation Index
> Fetch the complete documentation index at: https://docs.gulp.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Introduction

> Learn how to use the Osmosis Agent Improvement API to enhance your AI agents

<Note>
  **Early Access** - Osmosis API is currently in early access. [Contact us](mailto:founders@gulp.ai) to get started!
</Note>

# Welcome to Osmosis API

Osmosis Agent Improvement API helps you build smarter, more context-aware AI agents. Our API provides essential building blocks for creating agents that learn from past interactions and make better decisions over time.

<CardGroup cols={2}>
  <Card title="Quick Start" icon="bolt" href="/quickstart/authentication">
    Get up and running with Osmosis API in minutes
  </Card>

  <Card title="API Reference" icon="code" href="/api-reference/overview">
    Explore our comprehensive API documentation
  </Card>
</CardGroup>

## Key Features

<CardGroup cols={2}>
  <Card title="Contextual Enhancement" icon="brain">
    Enrich your agent's responses with relevant past knowledge using our powerful vector similarity search
  </Card>

  <Card title="Knowledge Storage" icon="database">
    Store and learn from interaction histories with efficient data management
  </Card>

  <Card title="Continuous Learning" icon="rotate">
    Improve agent responses based on past successes using our advanced learning algorithms
  </Card>

  <Card title="Smart Cleanup" icon="broom">
    Remove outdated or irrelevant knowledge automatically with intent-based deletion
  </Card>
</CardGroup>

## Core Endpoints

The API provides three main endpoints to power your AI agents:

<Steps>
  <Step title="/enhance_task">
    Enhance an agent's query with relevant knowledge from past interactions. Perfect for improving response quality and maintaining context.

    ```bash
    POST https://osmosis.gulp.dev/enhance_task
    ```
  </Step>

  <Step title="/store_knowledge">
    Store conversation or event history for future reference. Builds your knowledge base over time.

    ```bash
    POST https://osmosis.gulp.dev/store_knowledge
    ```
  </Step>

  <Step title="/delete_by_intent">
    Delete stale or irrelevant knowledge using intent matching. Keeps your knowledge base clean and relevant.

    ```bash
    POST https://osmosis.gulp.dev/delete_by_intent
    ```
  </Step>

  <Step title="/knowledge_status/{job_id}">
    Check the status of a knowledge upload job. Monitor progress and handle completion.

    ```bash
    GET https://osmosis.gulp.dev/knowledge_status/{job_id}
    ```
  </Step>
</Steps>

## Why Use Osmosis?

<Tip>
  AI Agent builders are often faced with the task of adding more and more edge case handling prompts to their agents.
  However, with osmosis it is now possible to attach those edge cases to the input prompt directly and keep the system prompts clean.
  We do this by using the [enhance\_task](/api-reference/enhance-task) endpoint, which will attach edge cases learned from past interactions stored via the [store\_knowledge](/api-reference/store-knowledge) endpoint.
</Tip>

Here's how Osmosis helps:

<AccordionGroup>
  <Accordion title="Knowledge Storage & Retrieval">
    * **Vector-Based Storage**: Efficiently store and retrieve interaction history
    * **Semantic Search**: Find relevant past experiences using advanced similarity matching
    * **Structured Data**: Maintain organized, queryable knowledge bases
  </Accordion>

  <Accordion title="Response Enhancement">
    * **Context Enrichment**: Add relevant past knowledge to current queries
    * **Pattern Recognition**: Learn from successful interactions
    * **Adaptive Responses**: Improve output quality over time
  </Accordion>

  <Accordion title="Knowledge Management">
    * **Smart Cleanup**: Automatically remove outdated information
    * **Intent Matching**: Delete irrelevant data based on semantic understanding
    * **Data Freshness**: Keep your knowledge base current and accurate
  </Accordion>
</AccordionGroup>

## Getting Started

To start using the Osmosis API:

<Steps>
  <Step title="Get API Access">
    [Contact us](mailto:founders@gulp.ai) to request API access
  </Step>

  <Step title="Authentication Setup">
    Follow our [authentication guide](/quickstart/authentication) to set up your API key
  </Step>

  <Step title="Make Your First Call">
    Try our [API endpoints](/api-reference/overview) with the provided examples
  </Step>
</Steps>

## Need Help?

<CardGroup cols={2}>
  <Card title="Email Support" icon="envelope">
    Contact us at [founders@gulp.ai](mailto:founders@gulp.ai)
  </Card>

  <Card title="API Reference" icon="book">
    Check our detailed [API documentation](/api-reference/overview)
  </Card>
</CardGroup>

Ready to get started? Head over to our [Quick Start guide](/quickstart/authentication) to begin!