Best tool for the job.
One verdict, the alternatives, and when to switch — not a list of forty tabs.
Claude Code
paidBest agentic coding tool right now — holds long multi-step tasks, edits across files, and runs its own checks before handing back.
Use it when: The default for working in real codebases.
Cursor
paidBest editor-native experience — inline diffs and tab-complete where you already work.
Pick this if: Choose if you live in an IDE and want chat + autocomplete fused.
aider
open sourceBest open-source route — a terminal pair-programmer that works with any model you bring.
Pick this if: Choose for OSS + bring-your-own-model.
Ollama
open sourceBest one-command local model runner — pull a model, start chatting.
Use it when: The default for local inference.
LM Studio
paidBest GUI for browsing and running local models.
Pick this if: Choose if you want a UI, not a terminal.
llama.cpp
open sourceBest low-level engine — the thing most local tools wrap.
Pick this if: Choose for max control or embedding into your own app.
MacWhisper
paidBest desktop transcription — Whisper quality, fully local, nothing leaves the machine.
Use it when: The default for private, on-device transcripts.
whisper.cpp
open sourceBest OSS engine — runs Whisper locally on basically anything, scriptable.
Pick this if: Choose to self-host or wire into a pipeline.
Otter.ai
paidBest for live meeting notes with speaker labels and search.
Pick this if: Choose for real-time team meetings.
Claude · Projects + web
paidBest for grounded, cited synthesis that pulls many sources into one brief.
Use it when: The default for multi-source briefs.
Elicit
paidBest for systematic literature review across academic papers.
Pick this if: Choose for evidence-grade / academic work.
GPT Researcher
open sourceBest OSS agent for autonomous web-research reports.
Pick this if: Choose to self-host a research agent.