多视角推理引擎 — 让 AI 从多个棱镜思考问题
Project description
pip install prismind
pip install 'prismind[web]' # 连带 Web 界面
为什么是 Prismind?
普通 AI 工具 = 问题进去 → 黑箱回答出来
Prismind = 问题进去 → 自动匹配分析框架 → 多视角拆解 → 辩论整合 → 结构化输出
内置 6 种分析视角(因果/辩论/对比/推演/历史/风险),自动适配模型能力(1.5B → GPT-4o)。
快速开始
from prismind import chat
result = chat("分析人工智能对医疗行业的影响",
model="deepseek-chat",
ak="sk-xxxxxxxx")
print(result)
模型自动适配
| 层级 | 典型模型 | 温度 | Token | 棱镜数 | 模式 |
|---|---|---|---|---|---|
| TINY | qwen2.5:1.5b | 0.3 | 2048 | 1 | 合并 |
| SMALL | qwen2.5:7b | 0.4 | 4096 | 2 | 串行 |
| FULL | deepseek-chat, gpt-3.5 | 0.5 | 8192 | 3 | 串行 |
| PREMIUM | gpt-4o, claude-3.5 | 0.6 | 16384 | 4 | 串行 |
三种模式
| 模式 | 说明 |
|---|---|
simple |
单次直答,最快 |
standard |
智能增强,短→直答 / 复杂→多视角串行+融合 |
deep |
多视角并行 + 辩论 + 综合 + 完整性检查(手动开启) |
三种接入
SDK
from prismind import chat, chat_stream, chat_async, set_mode, ImageLoader
for chunk in chat_stream("量子计算是什么?", ak="..."): print(chunk, end="")
result = await chat_async("分析市场趋势", ak="...")
img = ImageLoader.from_path("chart.jpg")
result = chat("分析这张图表", images=[img], ak="...")
CLI
prismind --ak sk-xxx --model deepseek-chat --mode standard
/mode simple|standard|deep 切换推理模式
/clear 清空会话
/tools 查看工具
Web
pip install 'prismind[web]'
prismind-web
# http://127.0.0.1:8081
配置优先级
运行时参数 > PRISMIND_AK 环境变量 > .env > prismind.json > 默认值
支持供应商
OpenAI / Ollama / Claude / Gemini / 百度 / 阿里 / 智谱
许可证
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distributions
No source distribution files available for this release.See tutorial on generating distribution archives.
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
prismind-0.2.0-py3-none-any.whl
(37.2 kB
view details)
File details
Details for the file prismind-0.2.0-py3-none-any.whl.
File metadata
- Download URL: prismind-0.2.0-py3-none-any.whl
- Upload date:
- Size: 37.2 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.12.10
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
85a32e277dbb6222fd3be1a1a114e11d311e7d4c4d978b034d732c6791499fd7
|
|
| MD5 |
fba3d2bf80ff5ba225ae8794614cceb2
|
|
| BLAKE2b-256 |
1c5e178d6c01df9ea73eca127dbc7cf9329ec26a06eb39fbbc87379f2820d52b
|