Updated
What makers build

AI search & research tools tech stack 2026: what 33 real products use

Answer engines, research assistants and deep-research tools that cite sources.

Among the 33 products we track in AI search & research tools, the usual stack is OpenAI API (LLM API, 85%), Next.js (frontend framework, 70%), Vercel (hosting, 58%) and LangChain (AI SDK & agent framework, 56%): shares of the products that show each decision.

How they're built

The typical stack

The leading pick where at least 8 of them show the decision and the leader has at least a quarter of it.

DecisionMost common pickShareAt default usage
LLM APIOpenAI API23 of 27 · 85%$10/mo · GPT-6 Luna
Frontend FrameworkNext.js16 of 23 · 70%—
HostingVercel11 of 19 · 58%$40/mo · Pro + Flat Rate CDN 10M
AI SDK & Agent FrameworkLangChain10 of 18 · 56%—
Backend FrameworkFastAPI12 of 15 · 80%—
DatabasePostgreSQL9 of 14 · 64%—
Transactional EmailResend7 of 12 · 58%$20/mo · Pro 50k
Document ParsingPyMuPDF4 of 11 · 36%—
MCP Server for Your ProductMCP SDK9 of 10 · 90%—
Vector Databasepgvector3 of 9 · 33%—
EmbeddingsSentence Transformers5 of 8 · 63%—
Open Model Hosting & GPUsHugging Face6 of 8 · 75%—

Monthly bills add up to about $70 at the calculators' default usage, list prices. Set your own usage →

Where they differ from other products

Tools AI search & research tools pick at least twice as often as products overall. Each bar is the share of products that show the decision and chose the tool.

What they chose, decision by decision

Among the AI search & research tools that show each choice, from makers' products and open-source code alike. Under each tool, some of the products that chose it.

LLM API27 products
OpenAI API23 · 85%Remy, Webhound, Cortex
Gemini API11 · 41%Webhound, Minicule, scira
Claude8 · 30%Sider, scira, MiroThinker
Groq3 · 11%scira, openscience, farfalle
Next.js16 · 70%Webhound, Cortex, Contrive
React11 · 48%scira, pipeshub-ai, SurfSense
Gradio3 · 13%MiroThinker, AutoRAG, OpenDeepSearch
Streamlit2 · 9%DeepResearchAgent, AutoRAG
Hosting19 products
Vercel11 · 58%Remy, Webhound, Cortex
GitHub Pages3 · 16%khoj, zotero-mcp, DeepResearchAgent
LangChain10 · 56%Minicule, pipeshub-ai, SurfSense
AI SDK5 · 28%scira, SurfSense, openscience
LlamaIndex4 · 22%sec-insights, rag-search, AutoRAG
LangGraph4 · 22%SurfSense, open-notebook, DeepResearchAgent
FastAPI12 · 80%pipeshub-ai, SurfSense, Cure AI
Hono2 · 13%openscience, xyne
Express2 · 13%pipeshub-ai, SurfSense
Flask2 · 13%MiroThinker, DeepResearchAgent
Database14 products
PostgreSQL9 · 64%scira, pipeshub-ai, SurfSense
Resend7 · 58%Cortex, Contrive, scira
Document Parsing11 products
PyMuPDF4 · 36%khoj, rag-search, zotero-mcp
MarkItDown3 · 27%MiroThinker, DeepResearchAgent, decipher-research-agent
pypdf3 · 27%SurfSense, sec-insights, rag-search
Docling2 · 18%pipeshub-ai, SurfSense
MCP SDK9 · 90%scira, pipeshub-ai, SurfSense
FastMCP4 · 40%pipeshub-ai, MiroThinker, zotero-mcp
Vector Database9 products
pgvector3 · 33%SurfSense, khoj, sec-insights
Qdrant2 · 22%pipeshub-ai, decipher-research-agent
Chroma1 · 11%zotero-mcp
Embeddings8 products
Sentence Transformers5 · 63%pipeshub-ai, SurfSense, khoj
Cohere3 · 38%scira, openscience, farfalle
Hugging Face6 · 75%Remy, MiroThinker, Cure AI
Together AI2 · 25%openscience, xyne
Modal1 · 13%openscience

Small samples, fewer than 8 products: Database Access & ORMs (SQLAlchemy 4, Prisma 2) · Payments (Stripe 6, Polar 1) · Search APIs for AI (Tavily 4, Serper 2) · AI UI Design (shadcn/ui 6) · LLM Observability & Evals (Langfuse 2, Arize Phoenix 2) · Code Sandboxes for Agents (E2B 3, Daytona 2) · Web Analytics (Plausible Analytics 2, Vercel Web Analytics 2) · Web Scraping for AI (Firecrawl 3, Crawl4AI 2) · Background Jobs & Cron (Celery 2, Trigger.dev 1) · File Storage (Amazon S3 3, Vercel Blob 1) · Product Analytics (PostHog 4) · Speech-to-Text & Text-to-Speech (Whisper 3, ElevenLabs 1) · Authentication (Better Auth 2, Clerk 1) · Error Monitoring (Sentry 3) · Image & Video Hosting (sharp 3)

AI search & research tools we track

23 makers' products and 10 open-source projects, makers' products first, each with the tools it uses (the most common here first).

Questions

What tech stack is behind AI search & research tools?

Among the 33 products we track in AI search & research tools, the usual stack is OpenAI API (LLM API, 85%), Next.js (frontend framework, 70%), Vercel (hosting, 58%) and LangChain (AI SDK & agent framework, 56%): shares of the products that show each decision. After those, FastAPI leads backend framework, PostgreSQL leads database and Resend leads transactional email.

Which tools are more common in AI search & research tools than elsewhere?

FastAPI for backend framework (2.1× as often as among products overall).

How much does a typical AI search & research tools stack cost?

About $70 a month in bills at the pricing calculators' default usage and list prices. The biggest lines: Vercel $40, Resend $20 and OpenAI API $10.

Which products in AI search & research tools are open source?

10 of the 33 are open-source projects, known from their public code, among them DeepResearchAgent, decipher-research-agent, Prismer, AutoRAG, farfalle and deep-research-web-ui.