AI chatbots tech stack 2026: what 41 real products use
General-purpose chat apps and chat interfaces over one or more models.
Among the 41 products we track in AI chatbots, the usual stack is React (frontend framework, 55%), OpenAI API (LLM API, 93%), LangChain (AI SDK & agent framework, 70%) and SQLite (database, 35%): 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.
| Decision | Most common pick | Share | At default usage |
|---|---|---|---|
| Frontend Framework | 17 of 31 · 55% | — | |
| LLM API | 25 of 27 · 93% | $10/mo · GPT-6 Luna | |
| AI SDK & Agent Framework | 14 of 20 · 70% | — | |
| Database | 6 of 17 · 35% | — | |
| Hosting | 13 of 17 · 76% | $40/mo · Pro + Flat Rate CDN 10M | |
| Vector Database | 6 of 16 · 38% | $51/mo · Starter | |
| Document Parsing | 5 of 14 · 36% | — | |
| Embeddings | 7 of 14 · 50% | — | |
| MCP Server for Your Product | 13 of 14 · 93% | — | |
| Run Models Locally | 10 of 12 · 83% | — | |
| Backend Framework | 6 of 11 · 55% | — | |
| Transactional Email | 5 of 11 · 45% | $20/mo · Essentials 50k | |
| Database Access & ORMs | 5 of 9 · 56% | — | |
| Open Model Hosting & GPUs | 7 of 9 · 78% | — | |
| Image & Video Hosting | 8 of 8 · 100% | — |
Monthly bills add up to about $121 at the calculators' default usage, list prices. Set your own usage →
Where they differ from other products
Tools AI chatbots 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 chatbots that show each choice, from makers' products and open-source code alike. Under each tool, some of the products that chose it.
Small samples, fewer than 8 products: AI UI Design (shadcn/ui 7) · Error Monitoring (Sentry 6) · Payments (Stripe 4, Lemon Squeezy 2) · Product Analytics (PostHog 5) · Search APIs for AI (Exa 2, Perplexity API 2) · Web Analytics (Plausible Analytics 2, Vercel Web Analytics 2) · Authentication (Better Auth 2, Clerk 1) · Mobile Framework (Expo 3, React Native 3) · AI Code Review (CodeRabbit 3) · Code Sandboxes for Agents (E2B 3) · File Storage (Amazon S3 2, Uppy 1) · LLM Observability & Evals (LangSmith 2, Langfuse 1) · Speech-to-Text & Text-to-Speech (ElevenLabs 1, AssemblyAI 1) · Web Scraping for AI (Firecrawl 2, Crawl4AI 1)
AI chatbots we track
19 makers' products and 22 open-source projects, makers' products first, each with the tools it uses (the most common here first).
Questions
What tech stack is behind AI chatbots?
Among the 41 products we track in AI chatbots, the usual stack is React (frontend framework, 55%), OpenAI API (LLM API, 93%), LangChain (AI SDK & agent framework, 70%) and SQLite (database, 35%): shares of the products that show each decision. After those, Vercel leads hosting, Chroma leads vector database and Unstructured leads document parsing.
Which tools are more common in AI chatbots than elsewhere?
Amazon Bedrock for LLM API (6.6× as often as among products overall), Mistral AI for LLM API (6.2× as often as among products overall) and Groq for LLM API (3.1× as often as among products overall).
How much does a typical AI chatbots stack cost?
About $121 a month in bills at the pricing calculators' default usage and list prices. The biggest lines: Chroma $51, Vercel $40 and SendGrid $20.
Which products in AI chatbots are open source?
22 of the 41 are open-source projects, known from their public code, among them NativeMindExtension, onyx-foss, cherry-studio, cherry-studio-app, py-gpt and LangBot.