Pricing you can act on, rate limits explained properly, git worktrees for parallel agents, and the reference tables the docs leave out. Written by the team that builds a workbench for these tools, and dated so you know how fresh each one is.
What the plans actually cost, what the rate limits actually mean, and the arithmetic for choosing between them.
All 20 pricing guides →Why you got cut off, when it comes back, and the settings that buy the most headroom.
Read guide →16 min Pricing, limits, and spendList prices are easy. This is what a session, a day, and a month actually come to, with the token math behind each.
Read guide →11 min Pricing, limits, and spendBoth seat prices, the per-member allowances behind them, and the arithmetic showing seats are not the governance premium everyone assumes.
Read guide →11 min Pricing, limits, and spendClaude Code pricing when the weekly cap fires: Pro vs Max vs API, two stacked subscriptions, or Get Plus.
Read guide →11 min Pricing, limits, and spendEleven changes ranked by what they actually save, with the two that are worth more than the other nine combined.
Read guide →9 min Pricing, limits, and spendOne allowance covers chat and Claude Code. Three separate caps enforce it, and they behave differently.
Read guide →9 minSeeing what you have spent and how much headroom is left, from the built-in command to full cost analytics by repo.
All 16 usage tracking guides →Seven real options, what each can genuinely see, and the one question that decides between them.
Read guide →8 min Usage tracking and analyticsThe JSONL files on your disk already record every token. Here is how to turn them into a number you can defend.
Read guide →8 min Usage tracking and analyticsFour metrics that change what you do, five that never will, a real benchmark to compare against, and how to read a skewed distribution.
Read guide →7 min Usage tracking and analyticsThe file format, the schema, the dedup rule, and the five traps that make naive implementations wrong.
Read guide →7 min Usage tracking and analyticsThree different numbers all get called usage. Which one you need depends on whether you are worried about being cut off or being billed.
Read guide →7 min Usage tracking and analyticsSeven categories of tool, what each can actually see, and the split that decides which one you need.
Read guide →7 minGetting the CLIs running on every platform, including the paths that are genuinely different on Windows.
All 19 install guides →Turn off the auto-updater first. Every downgrade that "did not work" skipped that step.
Read guide →6 min Install and setupSSH, mosh, Tailscale, browser sessions, Anthropic Remote Control, and a native iPhone fleet view, compared by where the agent actually runs.
Read guide →15 min Install and setupThe native installer replaced npm as the recommended path. Here is every route, who owns updates on each, and how to fix the common breakages.
Read guide →10 min Install and setupCommands, flags grouped by what they control, headless mode with its exact JSON result shape, and the background-session verbs most people never find.
Read guide →9 min Install and setupCloud sessions at claude.ai/code, how they hand off to and from your terminal, and the honest list of what a cloud VM cannot reach.
Read guide →9 min Install and setupFour routes, the one you should pick, and the PATH, shell, and filesystem problems that only happen on Windows.
Read guide →8 minConfigure and drive Claude Code properly: memory files, hooks, MCP, subagents, permissions, and headless runs.
All 51 reference guides →The engine reads your config; the GUI fights you. A profile in the shared config home is the part that genuinely works today.
Read guide →10 min Claude Code referenceThe desktop app has no custom-endpoint setting, but the same binary has a third-party inference mode that does — and Continuum speaks the wire it wants.
Read guide →9 min Claude Code referenceT3 Code drives CLIs rather than calling models itself, so you point the CLI, not T3 Code — through a provider’s environment-variable pane.
Read guide →8 min Claude Code referenceThe August 2026 lineup with real rates, every alias and what it resolves to, and a routing table you can apply per task.
Read guide →19 min Claude Code referenceThe complete ladder, with a label on every figure saying whether we read it from xAI or reconciled it from elsewhere.
Read guide →16 min Claude Code referenceThe top rung, priced against the two things that make it look cheaper than it is and the one thing that makes it look dearer.
Read guide →15 minConnect Claude Code to the tools and data it needs, choose servers worth trusting, and fix the failures that stop them loading.
All 9 mcp guides →The protocol, the server, and the two comparisons everyone gets wrong: MCP is not an API replacement and it is not RAG.
Read guide →10 min MCP for Claude CodeThree transport names, one of them deprecated, and a confusing amount of SSE still inside the transport that replaced it.
Read guide →9 min MCP for Claude CodeOne command to add it, the per-toolset URLs to keep it small, and the honest comparison against the gh CLI you already have.
Read guide →8 min MCP for Claude CodeThe endpoint that is actually current, why OAuth and API tokens cover different products, and the per-plan rate limits nobody mentions.
Read guide →8 min MCP for Claude CodeFourteen servers that earn their context, the exact install command for each, and the popular ones Claude Code already makes redundant.
Read guide →8 min MCP for Claude CodeA connector is an MCP server with a different front door. Which door you used decides where it lives, who controls it, and how you switch it off.
Read guide →8 minOpenAI's terminal agent: pricing, install on every platform, sandbox modes, config, and AGENTS.md.
All 22 codex guides →A seven-line block in config.toml points the Codex CLI at Continuum’s OpenAI-compatible API, with every model on your plan available by ID.
Read guide →8 min OpenAI Codex CLIInstall, auth, sandbox modes, config.toml, and the flags that actually change behaviour.
Read guide →13 min OpenAI Codex CLIFour model ids you can actually select today, six that were retired under you, and the rule for choosing between the live ones.
Read guide →10 min OpenAI Codex CLIOpenAI's official plugin makes Codex a callable worker inside Claude Code. Setup, rescue jobs, reviews, session transfer, and safe parallel work.
Read guide →7 min OpenAI Codex CLIIt stopped being a separate download in July 2026. Where it went, how to get it, and what it is actually for.
Read guide →7 min OpenAI Codex CLITwo orthogonal settings, the handful of pairings worth using, and the newer permission-profile system that replaces both.
Read guide →7 minInstall and run Google's terminal agent, understand its free and paid limits, and place Antigravity in the same working model.
All 5 gemini guides →Install it, work out which login still works after the June 2026 cutover, and drive it properly.
Read guide →10 min Gemini CLI & AntigravityThe agy binary, every flag it takes, the models it can reach, and what changes if you came from Gemini CLI.
Read guide →9 min Gemini CLI & AntigravityEvery install route, the Node version that actually matters, and the auth step that decides whether it works at all.
Read guide →7 min Gemini CLI & AntigravityTwo editions, four prices, and the June 2026 change that removed the free individual tier and the free Gemini CLI login.
Read guide →6 min Gemini CLI & AntigravityRemove the package, then remove the credentials and state that no package manager touches.
Read guide →5 minThe pricing model that has changed three times, what the free tier really allows, and how the agent works.
All 15 cursor guides →One API key and one base URL puts Claude, GPT, Grok, Gemini, and the free lane into Cursor’s model picker, billed to your Continuum plan.
Read guide →9 min CursorThe two pools, the dashboard that matters, what each action consumes, why old request counts are wrong, and how to track Cursor in a multi-agent stack.
Read guide →14 min CursorThe feature was renamed to Cloud Agents. Here is how it works, what it costs, where it breaks, and the cases where running agents locally still wins.
Read guide →14 min CursorWhy old Cursor prices contradict each other, what every major change meant in practice, and how to read the next announcement.
Read guide →13 min CursorCursor has several limits with different meters. Here is what each controls, when it resets, and the correct response at the ceiling.
Read guide →13 min CursorFive tiers, two separate usage pools, and one setting that decides whether a plan lasts a month or a week.
Read guide →9 minRunning several agents at once without them overwriting each other: the git mechanics and the failure modes.
All 16 worktrees guides →The operating model for a Claude Code fleet: isolated branches, bounded tasks, visible state, fresh review, and a merge queue one person can actually control.
Read guide →15 min git worktrees and parallel agentsThe operating system for parallel coding agents: one task, branch, worktree, owner, proof bundle, and review slot at a time.
Read guide →14 min git worktrees and parallel agentsOne repository, many working directories. The mechanics, the full command surface, and the six things that go wrong.
Read guide →10 min git worktrees and parallel agentsThe built-in --worktree flag, what it creates, how Claude Code enforces the isolation, and when to reach for git instead.
Read guide →10 min git worktrees and parallel agentsTwo to six agents at once: the isolation, the rate-limit math, and how to review the output without drowning.
Read guide →10 min git worktrees and parallel agentsThis is usually a workflow expansion, not a direct replacement. Continuum coordinates coding agents; Qodo reviews and governs the changes they produce.
Read guide →10 minWhat xAI’s always-on cloud agents actually are, what they cost, where the shared computer helps, and where it should worry you.
All 15 grok bot guides →A Bot can run <code>git clone</code>. Whether you should let it is a different question, and the answer turns on four things xAI has not shipped.
Read guide →12 min Grok BotFive numbers exist, all from one third-party source. The three numbers that decide your bill do not exist at all.
Read guide →11 min Grok BotThree ranking pages assert Linux support. The documentation says the opposite in one sentence. Here is the contradiction with its sources attached.
Read guide →7 min Grok BotSix eligible plans, three of them added on 21 August, an unpublished weekly allowance, and no spend cap behind any of it.
Read guide →20 min Grok BotA read of the documentation rather than the launch post, including the one claim the marketing pages and the docs do not agree on.
Read guide →17 min Grok BotThe page a buyer needs and nobody has written: the documented trust model, the missing certifications, and the questions to put to the vendor.
Read guide →16 minStraight comparisons between the tools you are actually choosing between. Continuum is not on either side of these.
All 70 head to head guides →These two get compared constantly and they are not the same kind of thing. One keeps working while you sleep; the other is the best pair programmer in a terminal. The interesting answer is not either.
Read guide →11 min Head to headEvery Grok Bot alternatives list currently ranking was written by one of the products on it. This one is not, and it says plainly where each option loses.
Read guide →17 min Head to headWhich AI is actually best for coding, split into the model that writes the code and the tool that runs it, with a pick for every kind of developer.
Read guide →17 min Head to headThe interface choices for one Claude Code session, several parallel agents, remote control, and serious diff review.
Read guide →14 min Head to headSix credible options, what each replaces, what it does not, and the one route that lets you keep using Cursor inside a different agent workbench.
Read guide →14 min Head to headFive workable shapes for parallel coding agents, from a phone-ready control plane to plain worktrees and tmux.
Read guide →14 minThe ideas underneath the tools: what agentic coding is, how to feed an agent context, and how to review what it wrote.
All 46 concepts guides →Grok Build resumes recent Claude Code work from its session UI, or imports a Claude transcript into a native Grok session. The handoff is real. Its fidelity is undocumented.
Read guide →8 min Agentic codingTwelve upgrades ordered by leverage, each with the payoff, the exact command or file, and the failure mode it prevents.
Read guide →17 min Agentic codingThe category got its vocabulary from marketing pages. This is the same category described by its mechanics, with an honest map of who each shape suits.
Read guide →16 min Agentic codingThe strongest case for the vendor VM, made properly, followed by the three costs that do not appear on the pricing page.
Read guide →15 min Agentic codingFour verdicts rather than one, because the tiers are four different products and only one of them is priced like a chatbot.
Read guide →15 min Agentic codingSix autonomous coding approaches ranked by accepted work, containment, context, reviewability, portability, and operating fit.
Read guide →15 minThe specific error strings, what actually causes each one, and the fix that holds.
All 25 troubleshooting guides →Filesystem denials, policy refusals, macOS folder protection, and the sandbox all look alike. How to tell which you have.
Read guide →8 min Errors and fixesThe exact checks to identify the empty pool, keep working safely, and prevent the same mid-task surprise next cycle.
Read guide →18 min Errors and fixesWhether it is Anthropic, your network, your proxy, or your credential, and how to tell in sixty seconds.
Read guide →14 min Errors and fixesWhich limit you hit, what to do in the next five minutes, and how to stop it recurring.
Read guide →14 min Errors and fixesTyping lag, slow Agent replies, high CPU, and Remote SSH stalls have different causes. This sequence separates them in minutes.
Read guide →12 min Errors and fixesMost Codex failures are the sandbox working correctly. How to tell those from real problems, starting with one command.
Read guide →11 minOne endpoint in front of every model provider: what a gateway buys you, which one to run, and how the market actually compares.
All 25 gateways guides →One API key in front of every model vendor. What it costs, what it hides, and the two cases where going direct is better.
Read guide →8 min LLM gateways and routingOne endpoint, many providers. What a gateway actually does, what it costs you, and the honest test for whether you need one at all.
Read guide →8 min LLM gateways and routingPortkey does four jobs at once, so there is no single replacement. Pick by which of the four you actually needed.
Read guide →7 min LLM gateways and routingThe enterprise one. Renamed, moved onto Unity Catalog, and now governing agents and MCP servers as well as model endpoints.
Read guide →7 min LLM gateways and routingNine gateways, seven columns, no vendor got to write its own row. Including the honest case for not running one.
Read guide →7 min LLM gateways and routingOne API key in front of hundreds of models, billed at provider list price. What it does well, and the three places it bites.
Read guide →7 minWhat AI tooling really costs an organization, and the controls that keep seats, keys, and token bills accountable.
All 18 spend guides →Nine things a team needs from an agent control plane, where each one lives in T3 Code, and the order they break in once you pass about eight engineers.
Read guide →15 min AI spend and governanceA procurement-shaped comparison of governed agent control planes, scored on the ten rows a security and finance review actually asks about.
Read guide →10 min AI spend and governanceThe five decisions an admin makes before the first developer runs the CLI, with the exact settings keys for each.
Read guide →10 min AI spend and governanceThe published $20 seat price, the metered usage sitting behind it, and the arithmetic against a Team Premium seat.
Read guide →9 min AI spend and governanceNine levers with the arithmetic attached, ordered by what they actually save rather than by how easy they are to write about.
Read guide →9 min AI spend and governanceUsage-based, per-token, invisible until the invoice. Four controls fix it, and only one of them is a hard limit.
Read guide →9 minThe companies that serve models over an API: what each one charges, where each is fast, and how to pick for coding work.
All 12 inference guides →Five kinds of company sell you inference. Here is what each actually charges, where each is genuinely fast, and which one a coding agent should point at.
Read guide →8 min Inference providersThroughput, first-token latency, and end-to-end turn time are three different numbers. Agents are bound by the one nobody advertises.
Read guide →8 min Inference providersCheap is only meaningful within a capability class. Here are the real numbers in each, and the five traps that turn a cheap rate into a large bill.
Read guide →8 min Inference providersSix providers that will serve you tokens for nothing, the exact limits, and what each one wants in return.
Read guide →7 min Inference providersFireworks serves open weights fast and sells the speed as a separate SKU. Here is what that buys, and where the catalog stops.
Read guide →7 min Inference providersBaseten is two products wearing one name: a small fast model API, and a serious platform for deploying your own model. Know which one you are buying.
Read guide →6 minContinuum runs Claude Code, Codex, Cursor, Gemini, and more under the subscriptions you already pay for, with live quota gauges and spend by repo.
free app · your subscriptions · local-first