← 返回
AI智能 中文

Pallio AI

Chat with Pallio AI knowledge-base personas. Ask questions against curated document collections with RAG-powered citations.
与 Pallio AI 知识库人物对话,针对精选文档集提问,获取 RAG 驱动的引用。
jayx-cloud
AI智能 clawhub v1.0.0 1 版本 100000 Key: 无需
★ 0
Stars
📥 435
下载
💾 6
安装
1
版本
#latest

概述

Pallio AI — Knowledge Base Chat

You can chat with a Pallio AI persona — an AI assistant backed by a curated knowledge base of uploaded documents. The persona answers questions using RAG (Retrieval-Augmented Generation) with document citations.

Setup

The user must set the PALLIO_PERSONA_ID environment variable to the ID of a public Pallio persona. Browse available personas at https://pallioai.com/community.

How to Use

Step 1: Initialize a Session

Before sending any messages, initialize a session to get an authentication token.

curl -s "https://pallioai.com/api/widget/init/$PALLIO_PERSONA_ID"

The response contains:

  • token — Session token (valid for 2 hours). Store this for all subsequent messages.
  • persona.name — The persona's display name.
  • persona.welcomeMessage — An introductory message from the persona. Show this to the user.
  • persona.starterPrompts — Suggested first questions. Offer these to the user.
  • persona.messageLimit — Maximum free messages per session (typically 3).

Step 2: Send Messages

Send the user's message with the session token. Maintain conversation history for context.

curl -s -X POST "https://pallioai.com/api/widget/chat" \
  -H "X-Widget-Token: $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "message": "What topics do you cover?",
    "history": []
  }'

The history array should contain all previous messages in the conversation:

[
  { "role": "user", "content": "What topics do you cover?" },
  { "role": "assistant", "content": "I cover fire safety procedures..." },
  { "role": "user", "content": "Tell me more about ventilation." }
]

The response contains:

  • response — The AI-generated answer. Display this to the user.
  • sources — Array of { title, page } objects. Display these as citations below the response.
  • messageNumber — Which message this is (1, 2, 3...).
  • messagesRemaining — How many free messages remain.
  • nudge — If "signup_soft", gently suggest signing up. If "signup_required", the session is exhausted.
  • signupUrl — URL for the user to create a full Pallio account.

Step 3: Display Sources

When sources is present and non-empty, format citations below the response:

Sources:
- Document Name (p. 42)
- Another Document (p. 15)

Step 4: Handle the Message Limit

Each session allows a limited number of free messages (typically 3).

  • When messagesRemaining reaches 0 or nudge is "signup_required", inform the user:

"You've used all free messages for this session. Sign up for full access: {signupUrl}"

  • To start a new session, call the init endpoint again (rate limited to 30/hour).

For unlimited access, the user can get a Pallio API key (Professional tier or higher) — see https://pallioai.com/settings.

Error Handling

Error CodeMeaningAction
----------------------------
SESSION_NOT_FOUNDToken is invalidRe-initialize the session
SESSION_EXPIREDToken TTL exceeded (2 hours)Re-initialize the session
MESSAGE_LIMIT_REACHEDFree messages exhaustedShow signup URL
RATE_LIMITEDToo many requestsWait and retry (check retryAfter field)
PERSONA_NOT_FOUNDInvalid persona IDVerify PALLIO_PERSONA_ID is correct
WIDGET_DISABLEDPersona owner disabled widgetTry a different persona

Important Notes

  • Sessions expire after 2 hours. If you get SESSION_EXPIRED, initialize a new session.
  • Rate limits: 30 session inits per IP per hour, 10 messages per IP per hour.
  • The persona only answers from its uploaded knowledge base — it will not fabricate answers.
  • This is the free widget endpoint. For full RAG pipeline access (hybrid search, more tokens, higher limits), use the Pallio Agent API with an API key.

版本历史

共 1 个版本

  • v1.0.0 当前
    2026-03-30 08:25 安全 安全

安全检测

腾讯云安全 (Keen)

安全,无风险
查看报告

腾讯云安全 (Sanbu)

安全,无风险
查看报告

🔗 相关推荐

ai-intelligence

ontology

oswalpalash
类型化知识图谱,用于结构化智能体记忆与可组合技能。支持创建/查询实体(人员、项目、任务、事件、文档)及关联...
★ 711 📥 243,706
ai-intelligence

Self-Improving + Proactive Agent

ivangdavila
自我反思+自我批评+自我学习+自组织记忆。智能体评估自身工作、发现错误并持续改进。
★ 1,355 📥 318,021
ai-intelligence

Proactive Agent

halthelobster
将AI智能体从任务执行者升级为主动预判需求、持续优化的智能伙伴。集成WAL协议、工作缓冲区、自主定时任务及实战验证模式。Hal Stack核心组件 🦞
★ 834 📥 212,990