Third-Party API

This page documents the API Key endpoints under the /tp group. They are designed for backend integrations, automation workflows, and external systems, and they cover document ingestion, section processing, knowledge graph, and content-generation flows.

Every endpoint on this page requires the API-KEY request header.

API keys can be created in Revornix via Avatar (bottom-left) -> Account -> APIKey Management.

Common Conventions

• All routes start with /tp.
• Most endpoints use POST with a JSON request body.
• /tp/file/upload uses multipart/form-data.
• /tp/document/create reads the optional X-User-Timezone header to calculate daily API collection limits and monthly summary dates in the user’s timezone.
• /tp/document/create always persists from_plat as api on the backend, so clients do not need to send it.

Integration flow sketch

Upload Endpoint

POST /tp/file/upload

Upload a file into the calling user’s default file system, then reference that path from /tp/document/create when creating a file document.

Parameters

Requirements

• The calling user must already have a default file system configured.
• Both file_path and content_type are required fields.

Example

import httpx
 
client = httpx.Client()
headers = {
    "API-KEY": "your api key"
}
files = {
    "file": ("example.md", open("example.md", "rb"), "text/markdown")
}
data = {
    "file_path": "/files/example.md",
    "content_type": "text/markdown"
}
 
response = client.post(
    "http://localhost:8001/tp/file/upload",
    headers=headers,
    files=files,
    data=data,
)

After upload, the usual next step is to call /tp/document/create and pass the uploaded file_path as the file document’s file_name.

Section Endpoints

Core write endpoints

Query and interaction endpoints

Comment endpoints

Label endpoints

POST /tp/section/create

Create a new section. This follows the same section processing model as the app, including trigger strategy, auto publish, auto podcast, and auto illustration settings.

Key fields

Example

import httpx
 
client = httpx.Client()
headers = {
    "API-KEY": "your api key"
}
 
response = client.post(
    "http://localhost:8001/tp/section/create",
    headers=headers,
    json={
        "title": "AI Product Watch",
        "description": "Track news and deeper analysis around AI products",
        "cover": "/covers/ai-weekly.png",
        "labels": [1, 2],
        "auto_publish": True,
        "auto_podcast": True,
        "auto_illustration": True,
        "process_task_trigger_type": 1
    },
)

POST /tp/section/ppt/generate

Generate PPT output for a section. This works well after the section Markdown is ready and you want presentation-ready slides.

Common field

POST /tp/section/comment/search

Returns paginated section comments, which works well for comment feeds and moderation tools.

Common fields

Document Endpoints

Label endpoints

Document write and processing endpoints

Document query and search endpoints

Note endpoints

POST /tp/document/create

Create a document using the same ingestion and processing flow as the app. The service always records from_plat as api.

Key fields

Category mapping

Requirements

• Clients do not need to send from_plat.
• The service records from_plat as api.
• On official deployments, this endpoint checks API document collection limits against the active plan.
• On official deployments, website and file documents may also hit extra entitlement checks after higher accumulated counts.
• If your workflow depends on day-based quota calculation, pass X-User-Timezone explicitly.

Example

import httpx
 
client = httpx.Client()
headers = {
    "API-KEY": "your api key",
    "X-User-Timezone": "Asia/Shanghai",
}
 
response = client.post(
    "http://localhost:8001/tp/document/create",
    headers=headers,
    json={
        "category": 1,
        "url": "https://example.com/article",
        "sections": [12],
        "labels": [3],
        "auto_summary": True,
        "auto_podcast": False,
        "auto_tag": True
    },
)

POST /tp/document/vector/search

Search the user’s knowledge vector store for the most relevant chunks to a question. This works well for retrieval augmentation and external question-answering orchestration.

POST /tp/document/note/search

Returns paginated note results with optional keyword filtering, which is useful for reading workspaces and annotation-style views.

Graph Endpoints

These endpoints return a unified graph response structure that can be used directly for knowledge graph visualization, node exploration, and graph browsers.