The mainstream AI stack — named, sorted, and linked.
No pay-to-play rankings: categories reflect how enterprises and prosumers actually buy software today. Expand a section for cards (three per row). Use the chips to isolate a category — everything stays collapsed until you open it.
Assistants & conversational chat Frontier models packaged for chat, copilot-style Q&A, and consumer-scale assistants.
ChatGPT
OpenAI’s general assistant: GPT-family reasoning, browsing/plugins (where enabled), voice, image understanding, and team workspaces for policy-controlled rollout.
Open site →Claude
Anthropic’s assistant focused on long context, careful instruction following, and “constitutional” safety framing — widely used for drafting, analysis, and code explanation.
Open site →Google Gemini
Google’s multimodal assistant tied to Workspace and Android surfaces; strong on Search grounding, Gmail/Docs side panels, and Android device-native “Gemini” experiences.
Open site →Microsoft Copilot
Microsoft’s consumer and work assistant across Edge, Windows, and Microsoft 365 — enterprise SKU ties Entra ID, Graph grounding, and compliance controls.
Open site →Meta AI
Meta’s assistant inside Facebook, Instagram, WhatsApp, and Ray-Ban Meta — optimized for social discovery, quick image generation, and on-device where available.
Open site →
Amazon Q
AWS’s assistant for builders and business users — IDE/CodeCatalyst integration, data/RAG connectors, and enterprise policy features aimed at regulated AWS shops.
Open site →Poe
Quora’s multi-bot chat hub — one subscription to hop between major models, bots, and custom prompts without juggling vendor accounts.
Open site →Le Chat (Mistral)
Mistral’s consumer assistant and document chat — strong EU footprint and open-weights-friendly positioning for teams diversifying off US-only stacks.
Open site →Grok
xAI’s model surfaced inside X — real-time-ish web flavor and “spicy” default tone; relevant when your users already live in the X graph.
Open site →DeepSeek
Reasoning-focused chat and APIs at aggressive price points — mainstream with developers for coding benchmarks and cost-sensitive R&D sandboxes.
Open site →Coding, agents & IDEs Autocomplete, chat-in-editor, and agent loops that edit repos — the fastest-moving buyer category in 2025–26.
GitHub Copilot
Microsoft/GitHub’s code completion and chat inside VS Code, JetBrains, and Neovim; Copilot Workspace targets multi-file tasks and PR-scale changes.
Open site →Cursor
Fork-of-VSCode editor built around agentic refactors, codebase-wide chat, and model pick-lists — popular with startups for shipping speed over toolchain purity.
Open site →Windsurf
Codeium’s agentic IDE (Cascade) emphasizing deep workspace awareness and flow-style edits — positioned as a Cursor-class competitor with strong free tier.
Open site →Replit
Browser IDE plus hosted runtimes and “Agent” workflows that scaffold apps, install deps, and deploy — mainstream for prototypes and teaching.
Open site →v0
Vercel’s UI generator from prompts — React/Tailwind/shadcn-style components you can iterate in chat then export into a Next.js repo.
Open site →JetBrains AI Assistant
AI features across IntelliJ, PyCharm, WebStorm, etc. — chat, commit messages, and refactor helpers aimed at teams already standardized on JetBrains IDEs.
Open site →Tabnine
Enterprise-friendly code completion with on-prem and air-gapped options — common where “send code to cloud LLM” is a security non-starter.
Open site →Sourcegraph Cody
Code-aware chat and search across large monorepos — shines when “find every caller of this symbol” matters as much as greenfield generation.
Open site →Codeium
Free-tier autocomplete and chat (same lineage as Windsurf) — default trial path for teams comparing Copilot alternatives at scale.
Open site →GitLab Duo
AI woven into MRs, security scans, and planning — buyer sweet spot is “we already run GitLab as the engineering system of record.”
Open site →Image, brand & product design Text-to-image, brand-safe marketing visuals, and UI design copilots — the default “creative suite” expansion path.
Midjourney
Discord- and web-based image generation with a distinctive aesthetic; widely used by designers and marketers for mood boards and campaign art.
Open site →Adobe Firefly
Adobe’s generative layer inside Creative Cloud — commercially indemnified workflows tied to Photoshop, Illustrator, Express, and enterprise asset libraries.
Open site →DALL·E
OpenAI’s text-to-image models surfaced in ChatGPT and APIs — strong prompt adherence; default choice when teams already standardize on OpenAI billing.
Open site →Canva Magic Studio
Template-first design suite with Magic Edit, generative video clips, and brand kits — dominant for marketing teams that prioritize speed over bespoke tooling.
Open site →Figma AI
First-screen UI generation, rename layers, translate copy, and search assets — AI sits inside the collaborative design source-of-truth most product teams already use.
Open site →Ideogram
Image model known for legible typography in posters and packaging mocks — common “second generator” beside Midjourney for text-in-image use cases.
Open site →Leonardo.Ai
Game and marketing asset pipelines with control nets, motion, and team libraries — popular when brand consistency beats one-off pretty prompts.
Open site →Stability AI
Stable Diffusion family and APIs — default path when you need weights-level control, fine-tunes, or self-host adjacent workflows.
Open site →Photoroom
E-commerce-first background removal and catalog shots — “marketplace seller” mainstream, not fine-art studio mainstream.
Open site →remove.bg
Single-purpose background removal that every marketing intern already knows — still the fastest path for one-off cutouts.
Open site →Video, audio & media Generative video, editing copilots, TTS, and creator-first pipelines — where GPU spend shows up in marketing orgs.
Runway
Gen-3 class video models, greenscreen, motion tracking, and collaborative timelines — reference-grade for short-form ads and concept reels.
Open site →Descript
Text-based video/audio editing, overdub voices, and filler-word removal — podcast and webinar teams treat the transcript as the primary timeline.
Open site →ElevenLabs
High-quality text-to-speech, voice cloning with consent workflows, and dubbing APIs — default audio layer for many SaaS demos and games.
Open site →CapCut
ByteDance’s mobile/desktop editor with auto-captions, templates, and generative helpers — dominant for short-form social pipelines outside the Adobe tax bracket.
Open site →Synthesia
Enterprise avatar videos from scripts — HR, L&D, and localization teams use it for compliant internal comms without booking studios.
Open site →Pika
Prompt-to-short-clip generation with strong social-native UX — common for rapid visual experiments alongside Runway in creative pods.
Open site →Loom
Async video messages with auto titles and summaries — “explain the deploy” mainstream inside distributed eng and GTM orgs.
Open site →Murf
Studio-style voiceovers for ads and e-learning — competes with ElevenLabs on project workflows more than raw cloning drama.
Open site →Otter.ai
Meeting transcription and summaries with calendar integrations — default stack for sales pods living in Zoom.
Open site →Adobe Podcast
Enhance Speech and mic checks for creators — quick cleanup when your “studio” is a closet next to a washer.
Open site →Work, docs & knowledge Where “AI features” ship inside the systems of record: docs, chat, calendars, and second-brain apps.
Notion AI
Summaries, Q&A across pages, database autofill, and meeting notes inside Notion — meets teams where the wiki already lives.
Open site →Microsoft 365 Copilot
Copilot in Word/Excel/PowerPoint/Outlook/Teams with Graph grounding — the enterprise bundle play for firms already on Entra + Purview.
Open site →Slack AI
Channel summaries, thread recap, and search answers constrained to workspace history — reduces scroll archaeology for big orgs on Slack.
Open site →NotebookLM
Google’s “sources-first” notebook: upload PDFs/docs, get grounded Q&A and audio overviews — popular with researchers avoiding generic web drift.
Open site →Coda AI
Doc-meets-database with AI columns, action buttons, and packs — ops teams use it for lightweight apps without shipping a full internal tool.
Open site →Google Workspace AI
Gemini-powered drafting in Gmail/Docs/Sheets/Meet — competes directly with Microsoft 365 for “AI inside the office suite” procurement bake-offs.
Open site →Linear
Issue tracking with AI issue intake and triage helpers — mainstream for product eng teams allergic to Jira’s default UX.
Open site →ClickUp Brain
All-in-one PM surface with Q&A across tasks, docs, and goals — wins RFPs where “single pane of glass” is non-negotiable.
Open site →Airtable AI
Structured data + automations with field-level AI — ops and marketing teams building lightweight apps without calling eng yet.
Open site →Atlassian Intelligence
AI across Confluence and Jira for summaries and writing — default for enterprises already paying Atlassian tax at planetary scale.
Open site →Search, answers & research RAG-native search UX — citations, PDF grounding, and “answer engines” replacing ten blue links for knowledge workers.
Perplexity
Conversational search with inline citations and follow-ups — mainstream default for “ChatGPT but with live web sources” among analysts and engineers.
Open site →You.com
AI modes for search, code, and writing with model pick-lists — positioned as privacy-aware alternative routing for teams avoiding single-vendor lock-in.
Open site →Brave Search / Leo
Independent index plus optional Leo assistant — appeals to security-conscious buyers who still want summarization without sending queries to Big Tech cores.
Open site →Consensus
Evidence-first answers over peer-reviewed literature — used in biotech, policy, and academia where “show me the papers” beats generic prose.
Open site →Elicit
Literature review workflows: extract variables from papers, build evidence tables, and track citations — common in research-heavy R&D orgs.
Open site →Google Search / AI Mode
AI Overviews and AI Mode in Search — widest reach by definition; buyers care about grounding, ad adjacency, and regional rollout differences.
Open site →Kagi
Paid search index with “Assistant” answers — mainstream among developers who want ranking without the ad stack shaping results.
Open site →Semantic Scholar
AI-sorted paper search, TL;DR snippets, and citation graphs — free layer under a lot of serious literature workflows.
Open site →Scholarcy
Flashcards and structured summaries from PDFs — speeds systematic reviews when paired with reference managers.
Open site →Phind
Developer-oriented answers with web + repo context — historically popular for “why is this stack trace happening” sessions (product direction may shift).
Open site →Automation & orchestration AI steps inside Zaps, scenarios, and webhooks — where line-of-business teams ship without waiting for engineering headcount.
Zapier AI
Natural-language Zap building, AI actions, and chatbots atop thousands of SaaS connectors — dominant iPaaS for SMB automation with AI on-ramps.
Open site →Make
Visual scenario builder (formerly Integromat) with OpenAI modules and iterators — popular when teams outgrow Zapier pricing but want low-code graphs.
Open site →n8n
Fair-code workflow engine you can self-host — engineering-led orgs pair it with private LLMs for ETL + LLM chains without SaaS data residency debates.
Open site →Bardeen
Browser-local automations with AI extraction — common for recruiting, sales prospecting, and scraping-adjacent tasks where users live in Chrome.
Open site →IFTTT AI
Consumer-scale applets plus AI services — less “enterprise iPaaS,” still mainstream for creators chaining social + smart-home + notifications.
Open site →Workato
Enterprise automation with recipes, governance, and embedded AI — typical for SAP/Salesforce/Workday stacks where IT needs audit trails.
Open site →Tray.io
Embedded iPaaS for SaaS vendors plus general automation — AI assists mapping fields and building connector logic for complex stacks.
Open site →Pipedream
Code-first workflows with thousands of app steps — engineers use it when Zapier’s UI is too tight but they still want speed.
Open site →Activepieces
Open-source automation with a Zapier-like builder — popular for self-host and EU data residency without giving up AI steps.
Open site →Power Automate
Microsoft’s cloud flows and RPA — default when Entra tenants already standardize on M365 + Dataverse + Copilot Studio adjacent glue.
Open site →APIs, models & developer infra Where products get built: hosted inference, weights hubs, and fast LPU/GPU clouds behind your own prompts and evals.
OpenAI API
GPT family, embeddings, realtime voice, and Assistants/Batches APIs — default vendor for startups needing broad model coverage and predictable SLAs.
Open site →Anthropic API
Claude models with long context and tool use — common “second provider” for red-team diversity, coding agents, and safety-sensitive enterprises.
Open site →Google AI Studio
Gemini API keys, prompt playground, and tuning flows — low-friction on-ramp for builders already in GCP billing and Vertex migration paths.
Open site →Hugging Face
Model hub, Spaces demos, datasets, and Inference Endpoints — lingua franca for open-weights workflows and academic reproducibility.
Open site →GroqCloud
LPU-backed inference marketed for very low latency tokens — mainstream for voice agents, streaming UIs, and eval harnesses that stress TTFT.
Open site →Replicate
One-line HTTP APIs over community models — image/video/audio microservices without owning GPUs; common in hackathons and indie SaaS MVPs.
Open site →Together AI
Hosted open-weight inference and fine-tuning — common “second cloud” for Llama-class models and dedicated GPU clusters.
Open site →Cohere
Embeddings, rerank, and command models with enterprise controls — strong where retrieval + RAG quality beats raw chat novelty.
Open site →Mistral API
European-headquartered model APIs with competitive pricing — procurement pick when data residency and vendor diversification matter.
Open site →Fireworks AI
Fast inference for open models and compound AI systems — shows up in prod stacks that need low-latency OSS endpoints.
Open site →Modal
Python-first serverless GPU jobs — mainstream for batch inference, fine-tunes, and internal tools without owning k8s GPU fleets.
Open site →Logos via public favicon service for identification only; trademarks belong to their owners. AI Donkey Chase does not endorse vendors — this page is an orientation map, not a ranking.