Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/thenoname-gurl/EcliPanel/llms.txt

Use this file to discover all available pages before exploring further.

EcliPanel ships two AI surfaces — AI Chat for conversational interactions and AI Studio for more advanced prompting — backed by a model management layer that lets admins configure which AI models are available and restrict access by user or organisation. All AI features are controlled by the ai feature flag. If your deployment does not have AI configured, these endpoints and the AI navigation items are hidden entirely.
AI features require the ai feature flag to be enabled in your panel settings. Without it, all /api/ai/* endpoints return a feature-disabled error and the AI sections do not appear in the dashboard navigation.

​
AI Chat

AI Chat is a lightweight conversational interface available to all users (no tier restriction): 
POST /api/ai/chat
It is accessible at /dashboard/ai-chat in the frontend. The endpoint accepts a message and returns a streamed or buffered completion from the configured model.

​
AI Studio

AI Studio provides a more capable interface for advanced prompting, code generation, and longer sessions. It is restricted to Pro (paid) tier and above: 
POST /api/ai/studio
The Studio UI is at /dashboard/ai-studio and requires the paid portal tier in addition to the ai feature flag.

​
OpenAI-compatible endpoint

EcliPanel exposes a drop-in OpenAI-compatible chat completions endpoint that you can point any OpenAI SDK or HTTP client at: 
POST /api/ai/openai/v1/chat/completions
POST /api/ai/openai/v1/completions
This means you can use the official openai Python or Node.js library with EcliPanel as the base URL: 
from openai import OpenAI

client = OpenAI(
    base_url="https://your-panel.example/api/ai/openai/v1",
    api_key="<your-eclipanel-api-key>",
)

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "How do I restart my Minecraft server?"}],
)
print(response.choices[0].message.content)
The backend proxies the request to the configured upstream model endpoint using requestWithFallback, which automatically retries against alternative endpoints if the primary one is rate-limited.

​
Model management

Admins configure which upstream models are available to the panel. Each AIModel record stores one or more endpoint definitions (base URL + API key). The backend cycles through endpoints and applies per-endpoint cooldowns when rate limits are encountered: 
GET  /api/ai/models             # list models available to me
GET  /api/ai/my-models          # my personal model access list
GET  /api/admin/ai/models       # admin: list all models
POST /api/admin/ai/models       # admin: register a new model

​
Per-user and per-org model access

Admins can grant specific models to individual users (AIModelUser) or to entire organisations (AIModelOrg). When the backend resolves available models for a request, it checks both the user’s direct grants and the grants for all organisations they belong to: 
POST /api/admin/users/:id/ai/:linkId   # link/update a model grant for a user
Users without an explicit grant can only access models that have been made globally available.

​
Usage tracking

Every AI request is recorded in the AIUsage table. This lets admins audit consumption per user and per organisation through the SOC dashboard: 
GET /api/soc/usage/user/:id    # AI usage for a specific user
GET /api/soc/usage/org/:id     # AI usage for a specific organisation
Rate-limit events on upstream AI endpoints are also logged to Redis (admin:ai:cooldowns) for admin visibility.

Build docs developers (and LLMs) love

Get started for freeTalk to us