The Taku Blog
Ideas worth building on.
Stories, product notes, and perspectives on how AI gets discovered, used, remixed, and shared.
October 1, 2026
ElevenLabs Alternatives: Studios, APIs, Open Source
The best ElevenLabs alternatives by job: voiceover studios, low-latency voice APIs, big-cloud TTS, and free open-source models you can run yourself.

October 1, 2026
Gemini CLI vs Claude Code, and Antigravity CLI
Gemini CLI vs Claude Code after Google moved free and Pro users to Antigravity CLI. Install, models, context, extensibility, pricing model and licence.

October 1, 2026
HeyGen Alternatives: 8 AI Avatar Tools Compared
The best HeyGen alternatives for avatar videos, training content, translation and editing, plus HeyGen vs Synthesia and what consent rules to expect.

October 1, 2026
Jasper AI Alternatives, Sorted by What You Need
The best Jasper AI alternatives in 2026, from free writing tools to SEO and sales platforms, plus which old 'AI copywriters' have quietly changed category.

October 1, 2026
Midjourney Alternatives: 10 Picks Sorted by Job
The best Midjourney alternatives in 2026, grouped by what you need: free tools, no Discord, open-source FLUX and Stable Diffusion, text and vectors.

October 1, 2026
Otter.ai Alternatives: AI Note Takers Compared
The best Otter.ai alternatives for meeting notes: Fireflies, Fathom, tl;dv, Granola and more, plus the AI notes already built into Zoom, Teams and Meet.

October 1, 2026
Retool Alternatives: Sorted by Licence and Hosting
Open source, self-hosted, code-first or AI-first: how Appsmith, Budibase, ToolJet, Refine, Power Apps and others compare as Retool alternatives.

September 30, 2026
Claude Code Hooks: Events, Matchers, and Examples
Claude Code hooks run your own commands at fixed points in a session. The events, where hook settings live, how matchers work, and how a hook blocks.

September 30, 2026
Claude Code Subagents: How They Work and When to Use Them
Claude Code subagents run side tasks in their own context window. Where the files live, the frontmatter fields, and when a skill or agent team fits better.

September 30, 2026
CLAUDE.md File: What to Put in It and Where It Goes
What a CLAUDE.md file is, where Claude Code looks for it, what to put in it, how @imports work, and how it relates to AGENTS.md.

September 30, 2026
Claude Projects vs Custom GPTs: Which to Build On Now
Claude Projects are a private workspace built around your files. Custom GPTs are shareable assistants that OpenAI is now moving to plugins. How to choose.

September 30, 2026
Claude Skills for Developers: 8 Worth Installing, by Job
The Claude skills for developers worth installing, grouped by the job they do: commits, SQL, browser testing, AGENTS.md, prototypes, and diagrams.

September 30, 2026
Cursor Rules: Project Rules, .mdc Files, and AGENTS.md
How Cursor rules work in 2026: the four rule types, the .cursor/rules .mdc format, the four apply modes, and how to migrate off the legacy .cursorrules file.

September 30, 2026
Cursor vs VS Code: Is the AI Fork Still Worth Switching To?
VS Code now ships chat, agents, MCP and bring-your-own-model support. Here is what Cursor still does better, and when staying in VS Code is the smarter call.

September 30, 2026
How to Build an MCP Server in Python or TypeScript
Build an MCP server with the official v2 SDKs: define a tool, serve it over stdio, test it in the MCP Inspector, and connect it to Claude Code.

September 30, 2026
Langfuse vs LangSmith: Open Source vs the LangChain Platform
Langfuse vs LangSmith compared on licence, self-hosting, tracing, evals, prompts, deployment and billing model, so you can pick an LLM observability tool.

September 30, 2026
LangGraph vs LangChain: Which Layer Do You Actually Need?
LangChain's agents now run on LangGraph. Here's what each layer does, when to drop down to LangGraph, and whether you need LangChain at all.

September 30, 2026
MCP vs API: The Difference, and Where Function Calling Fits
An API is a service's own interface. MCP is a standard way for AI apps to discover and call tools that usually wrap APIs. Plus where function calling fits.

September 30, 2026
Ollama vs llama.cpp: Same Engine, Different Job
Ollama builds on llama.cpp, so the real choice is who makes the decisions: Ollama's defaults or your own llama-server flags. How to pick.

September 30, 2026
OpenAI Agents SDK vs Claude Agent SDK: Library vs Harness
OpenAI's Agents SDK is a light multi-agent library for any model. Claude's Agent SDK is Claude Code's full agent loop. Here's which one fits your build.

September 30, 2026
OpenRouter Alternatives: Pick by Why You're Leaving
OpenRouter alternatives sorted by the reason you're switching: self-hosted control, a gateway in front of your own keys, one cloud contract, or local models.

September 30, 2026
Relevance AI Alternatives: 6 Options, Sorted by Why You're Leaving
Six Relevance AI alternatives grouped by the reason you're looking: a single assistant, self-hosting, a developer framework, or your existing suite.

September 30, 2026
vLLM vs Ollama: A Serving Engine and a Local Runner
vLLM is built to serve many users from GPUs; Ollama is built to run models on your own machine. What that means for setup, concurrency and production.

September 29, 2026
Claude Code Alternatives: Terminal Agents Sorted by Why You're Leaving
Leaving Claude Code over cost, model lock-in, licensing, or local models? Each reason points to a different terminal agent. What each one actually offers.
