AI Services in Search Engine & RAG

AI search engines that leverage semantic understanding and RAG technology for information retrievalFeaturing 5 AI services in total.

パープレキシティ

Perplexity

An answer-first AI search engine that cites source URLs alongside every response. Supports multiple LLMs including GPT, Claude, Gemini, and Sonar, with file upload and image generation available on Pro plans. Offers Perplexity Pages for generating summary articles and an AI assistant experience — a rapidly growing service redefining how people search.

# 検索・レコメンド# 検索エンジン・RAG# AIエージェント# 要約・翻訳# 生成AI（LLM）# 日本語対応

メモリーク

Memoriq

A private AI memory vault that lets you save and search conversations from ChatGPT, Claude, Gemini, and Grok in your own end-to-end encrypted (E2EE) storage. A browser extension automatically captures conversations and encrypts them on your device before upload. Open-source and self-hostable. Keeps your valuable AI interactions intact without entrusting plaintext to the cloud.

# 検索エンジン・RAG# AIエージェント# 業務効率化# 生産性向上

ヴァンナエーアイ

Vanna AI

An open-source (MIT) RAG framework for Text-to-SQL that generates accurate SQL from natural language questions. Train it on your database DDL, documentation, and SQL examples, and it uses Agentic Retrieval to produce high-precision queries. Supports major databases including PostgreSQL, Snowflake, and BigQuery, as well as leading LLMs from OpenAI, Anthropic, and others.

# 検索エンジン・RAG# コード生成# AIエージェント# 生成AI（LLM）

カサナレ

Kasanare

A business-specific generative AI product built to fit each company's unique operations. It combines RAG to teach the AI domain-specific terminology and workflows, usage governance aligned with organizational policies, and action controls that operate existing systems — enabling AI agents that go beyond answer generation to complete real work end-to-end. Deployed at organizations including Mitsubishi UFJ Trust and Banking. Developed by Kasanare Inc.

# AIエージェント# AIチャットボット# 検索エンジン・RAG# 業務効率化# 自動化・省力化# 日本語対応

ディープシーク

DeepSeek

A series of large language models (LLMs) developed by Chinese AI company DeepSeek. The open-source DeepSeek-R1 attracted wide attention for matching or surpassing GPT-4o and Claude 3.5 Sonnet on reasoning benchmarks. Known for an efficient Mixture of Experts (MoE) architecture and open weights available for commercial use.

# 検索エンジン・RAG# 生成AI（LLM）