API Documentation v7.22

The PLI 7 API provides programmatic access to a hybrid AI engine trained exclusively by Seridarivus. Integrate real-time research, verified media retrieval, and advanced UI reasoning into your own apps.

Authentication

Access to the API is controlled via API Keys. All requests must include the x-api-key HTTP header.

x-api-key: pli7_your_secret_key_here

Note: You can manage and revoke your keys in the PLI Studio (https://plistudio.net) dashboard settings.

Chat / Reasoning

POST https://api.plistudio.net/v1/chat

Sends a prompt to the AI and triggers a sequence of research tasks based on your selected sources.

Parameter	Type	Description
message	string	Required. The user query or instruction.
sources	array	Optional. List of sources to search (e.g., `["web", "wiki", "youtube", "pexels"]`).
context	object	Optional. Metadata like `localTime` for temporal reasoning.

Example Payload

{
  "message": "Explain the current status of the James Webb Telescope",
  "sources":["web", "wiki"],
  "context": { "localTime": "2026-04-26 11:57 PM" }
}

Available Data Sources

When sending a request, use the sources array to enable specific research modules.

wiki: Fetches encyclopedic context and historical data.
web: Scans the live internet for recent news and specific data points.
youtube: Identifies relevant video content and documentaries.
pexels: Retrieves high-resolution stock photography.

Research & Citations

If a query requires data, PLI 7 uses an internal search tool and provides explicit citations for transparency.

Citation Format

The AI will inject a <cite> tag immediately following the information it supports.

<cite name="Source Name" url="https://link.com"/>

Frontend Action: Parse this tag as a clickable superscript or a bibliographic link. Do not display the raw tag to the user.

Parsing Annotations & Tags

The PLI 7 API returns a rich, proprietary syntax designed to trigger dynamic UI components. You must parse these tags on your frontend to render the hybrid experience.

Example: Image Generation Output

When a user asks the AI to generate or draw an image, the API returns a response containing a high-fidelity prompt inside <plii4> tags.

{
    "answer": "I have visualized the dragon in its alien environment for you. <nIMG<>>\n\n<plii4>A majestic dragon soaring through a vibrant, alien sky. The dragon's scales shimmer with iridescent hues of emerald, sapphire, and amethyst, reflecting the light of two large, ringed moons hanging in the celestial expanse. Its wings, vast and membranous, are patterned with intricate, bioluminescent veins that pulse with a soft, golden glow. Below, a surreal landscape unfolds: crystalline trees with glowing foliage reach towards the heavens, their branches adorned with delicate, bell-shaped flowers that chime in an unseen breeze. Rivers of liquid silver flow through valleys carpeted with phosphorescent moss, illuminating the twilight world. In the distance, jagged, obsidian mountains pierce the horizon, their peaks wreathed in swirling nebulae of violet and cyan. The overall atmosphere is one of mystical wonder and untamed, alien beauty.</plii4>",
    "media":[],
    "sources_used":["web"],
    "timestamp": "2026-04-26T23:57:08.691Z"
}

Tag Reference Guide

Tag / Notation	Purpose	How to handle it
<cite name="..." url="..."/>	Research Citation	Render as a clickable link or superscript (e.g., [Source Name]).
<plii4>...</plii4>	Image Generation	Contains a ~350-word descriptive prompt. Hide this text and pass it to your image generation pipeline (e.g., PLII4.2U model).
<nIMG<>>	Media Suppressor	Present when AI generates a custom visual. Remove this tag and hide external stock images/Pexels from the UI.
<map>Location</map>	Geolocation Trigger	Render a button that opens Google Maps or an embedded map iframe for the specified location.
<diagram>...</diagram>	Visual Engine	Contains Mermaid syntax. Pass to a Mermaid.js renderer to display flowcharts/graphs.
<stp>...</stp>	Live UI Sandbox	Contains raw HTML/CSS/JS. Render inside a sandboxed `<iframe>`.
</Keyword/>	Visual Media Focus	Highlights the primary visual subject. Use for keyword bolding or media fetching.

Handling Standard Formatting

PLI 7 uses standard Markdown for basic text. You should parse **bold**, *italics*, and * bullets.

STRICT CONSTRAINT: The engine will never use acute accents (é, á, ó) in text or code.