Conan
Conan is a native macOS HUD for Claude Code that streams every prompt, tool call, skill and token live as the session runs.
If you live in Claude Code and want receipts on how the agent spends tokens and which skills actually fire, Conan earns its $29 once. The Context breakdown and the Timeline's 'skills considered' rows remain the clearest agent observability we've seen at this price, and v1.4.0's inline subagent activity plus the Project skills view close two real gaps. Just know the lane is narrow: Claude Code, macOS 13+ on Apple silicon, no subscription, no Windows.
Verified 3d ago · liveness 77/100 · cite: rightaichoice.com/tools/conan
- Developers running Claude Code for complex multi-skill sessions
- Teams auditing AI agent token consumption and context efficiency
- Prompt engineers needing visibility into skill selection and tool calls
- Power users managing context limits in long-running coding sessions
- Non-developers or anyone not using Claude Code
- Windows or Linux users (macOS 13+ Apple silicon only)
- Teams looking for a Claude Code replacement rather than an observer
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Conan if you don't run Claude Code, you're on Windows, Linux, or an Intel Mac, or you want a bare terminal without observability panels overlaid on it.
The free tier is observe-only — Context, Pulse, Skills & MCP, Files and Agents all need the $29 Premium unlock to become fully useful for auditing.
At $29 once for the Premium unlock (lifetime of 1.x, unlimited Macs), Conan sits well below subscription-priced developer tooling — it roughly matches two months of a typical $15/mo observability add-on. It fits solo developers and small teams already paying for Claude Code who want token and skill visibility without a recurring line item. If you need Windows or Linux coverage, no amount of budget solves it today.
In short
Conan — Conan is a native macOS HUD for Claude Code that streams every prompt, tool call, skill and token live as the session runs. Best for Developers running Claude Code for complex multi-skill sessions, Teams auditing AI agent token consumption and context efficiency, Prompt engineers needing visibility into skill selection and tool calls. Free to start; paid plans from $29.
What's new in Conan
Checked yesterdayAcross the latest 5 updates: 4 feature updates and 1 launch.
v1.4.0 — Agent activity in the Timeline
Subagent spawns now show inline in the Timeline with live status, duration, model and tools. The Skills tab gains a Project view, and usage windows stay fresh via background polling with a live/fallback badge.
v1.3.0 — Agents in the HUD
A new Agents tab lists every Claude Code subagent installed on your machine — global, project and plugin-shipped — with description, tools and on-disk path. YAML block-scalar skill descriptions now render correctly.
v1.2.0 — Conan Setup in the box
New terminals can open into a bundled spec-first setup that interviews you and starts Claude with a reviewed PRD.md and CLAUDE.md. Auto context tracking is now opt-in.
v1.1.0 — Files panel
A Files panel browses the active terminal's working directory beside the Timeline and follows your cd, flagging files Claude has read or edited this session.
Conan 1.0 — first public release
A terminal-primary native macOS app that wraps and observes Claude Code, with the Context, Usage, Pulse, Skills and MCP HUD plus Claude Radio. Free as a live observer; Premium $29 once unlocks the full insight layer.
What people actually say about Conan — is it worth it?
We scanned public community sources for Conan on Aug 21, 2026 and could not establish that the discussion we found is about this tool rather than something else sharing its name. Our own analysis of that scan says the posts were off-subject. Rather than publish a sentiment score built on the wrong subject, we publish nothing here and re-run the scan.
Viability Score
How well maintained and how widely used is Conan? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- Live HUD overlaying Claude Code on macOS
- Streaming Timeline of prompts, hooks, tool calls and skills
- Subagent spawns inline in Timeline with status, duration, model and tools (v1.4.0)
- Context window breakdown: system, tools, memory, skills and messages tokens
- Pulse throughput chart showing when Claude works and when it waits
- Skills & MCP inventory with last-fired timestamps and on-disk paths
- Project skills view alongside User and System views (v1.4.0)
- Agents tab listing global, project and plugin-shipped subagents (v1.3.0)
- Files panel following cd and flagging files Claude read or edited (v1.1.0)
- Conan Setup spec-first flow producing PRD.md and CLAUDE.md (v1.2.0)
- Claude Radio with curated stations and any YouTube URL, session-only
- Background usage polling with live/fallback badge (v1.4.0)
- Live /usage read feeding Context and Pulse panels
- Native macOS terminal running your normal Claude Code login
About Conan
Conan is a native macOS app that wraps Claude Code in a live HUD, so an agent session becomes something you watch instead of something you squint at in a scrolling terminal. You keep running the real Claude Code in its own terminal with your normal login; Conan lays observability panels alongside it. It's built for developers already living in Claude Code who want to know where tokens go and which skills actually fire — and for anyone auditing context efficiency across long, multi-skill sessions. The Timeline streams every prompt, hook, tool call and skill as it happens, and as of v1.4.0 subagent spawns appear inline with live status, duration, model and tools rather than as a black box. The Context panel breaks the window into system, tools, memory, skills and messages tokens against a live /usage read, while the Pulse chart spikes when Claude works and goes quiet when it waits. A Skills & MCP inventory lists what's in play across User, System and Project views with last-fired timestamps and on-disk paths, the Files panel follows your cd and flags files Claude read or edited, and the Agents tab (v1.3.0) lists global, project and plugin-shipped subagents. Setup can be spec-first if you want it: new terminals can open into Conan Setup (bundled since v1.2.0), which interviews you and produces a reviewed PRD.md and CLAUDE.md before Claude starts — skippable with one keystroke. Claude Radio plays curated stations like Claude FM or any YouTube URL, session-only. It requires macOS 13+ on Apple silicon; the page notes Windows and Linux users can only join a notification list. Free to download as a live observer, Premium $29 once unlocks the full insight layer with no subscription. Positioned against raw terminal output and generic tracing tools, Conan is narrower by design — it observes Claude Code rather than replacing it, and it won't help you if you're not already in Claude Code on a Mac. Within that lane, the context breakdown is the part competitors don't match.
Behind the Verdict
Reach for Conan when you're already deep in Claude Code and the question you keep asking is 'why did that turn cost so much?' The Context panel answers it directly — system, tools, memory, skills and messages tokens split against a live /usage read — and the Timeline's considered-vs-fired skill rows tell you whether your skill descriptions are actually doing work or just sitting on disk. That second one is the feature we use most. We'd also pick it up in multi-skill or subagent-heavy sessions. Before v1.4.0, a subagent spawn was effectively invisible; now it shows inline with status, duration, model and tools, and the Project view means project-scoped skills stop hiding behind the global list. For prompt engineers tuning skill selection, that's the difference between guessing and measuring. Pass on it if you're on Windows or Linux. The vendor page is explicit — those users get a notification list, not a build. Even on a Mac, you need Apple silicon and macOS 13+, so an older Intel laptop is out. And if your Claude Code usage is short, single-turn, and cheap, the HUD is overhead you'll stop opening. The closest alternative is just the terminal itself. Claude Code's own output plus /usage gets you part of the way for free, and if you're comfortable reading scrolling logs you may not feel the $29. Conan's advantage is structure — the same data arranged so a 1.0M-token window's composition is obvious at a glance rather than something you reconstruct. One caveat on the observer model: Conan wraps Claude Code, it doesn't replace it, so your workflow stays terminal-primary and the HUD is a second surface to watch. Dismissing it costs nothing, but if you want the observability you have to actively look at it. Claude Radio is a genuine nice-to-have rather than a reason to
Researching Conan? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Conan actually fits — and what changes day-one when you adopt it.
You open a new terminal into Conan Setup, answer a few questions, and get a reviewed PRD.md and CLAUDE.md before Claude starts. As the session runs, the Timeline streams prompts, hooks and skills while the Context panel shows the window filling toward 1.0M tokens.
Outcome: You spot the context pressure before it truncates the session, and you finish the task with a handoff-ready project spec rather than a half-remembered terminal scrollback.
A skill you expect Claude to fire isn't firing, so you open the Timeline and read the 'skills considered' rows — including partial and strong description matches against the terms in your prompt — alongside the Skills & MCP inventory's last-fired timestamps.
Outcome: You rewrite the skill description with evidence of what Claude is actually matching on, instead of guessing why the skill got skipped.
Across a multi-subagent build, the Timeline shows each subagent spawn inline with status, duration, model and tools, the Pulse chart shows throughput spikes, and the Context panel breaks tokens into system, tools, memory, skills and messages. Live /usage reads keep the numbers fresh via background polling.
Outcome: You can attribute token spend to specific skills and subagents, and decide which ones to trim before the next sprint rather than after the bill.
Use Cases
- Watch the context window fill in real time to avoid hitting limits mid-session
- Debug why Claude picks one skill over another using the 'skills considered' heuristic
- Track session throughput to distinguish active work from a waiting Claude
- Review a full timeline of tool calls and prompts after a session for audit or replanning
- Run ambient focus audio via Claude Radio while coding with Claude Code
- Start a new project with Conan Setup to generate a PRD.md and CLAUDE.md before Claude runs
- Track subagent spawns and tool usage across long sessions for performance tuning
- Follow file reads and edits via the Files panel to understand filesystem activity
Models Under the Hood
as of 2026-10-17
Limitations
- Conan is a native macOS app requiring macOS 13+ on Apple silicon — there is no Windows or Linux build, only a notification list.
- It wraps Claude Code and does not replace it: you need Claude Code and a working Claude login to use it.
- It is free to run as a live observer; the full insight layer requires a one-time $29 Premium purchase.
- Auto context tracking is opt-in, so context measurement is off unless you enable it.
as of 2026-10-09
Verification history
We have re-verified Conan 10 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-checked, vendor evidence unchanged
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 10 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Conan tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Free
$0
Ideal for
Developers who want to try Conan as a live observer around their existing Claude Code terminal before spending anything.
What this tier adds
Free entry point: download and run as a live observer with the Timeline and Claude Radio, without the full insight layer.
Premium
$29 one-time
Ideal for
Developers running long multi-skill Claude Code sessions who need the Context breakdown, Pulse, Skills & MCP inventory, Files and Agents panels.
What this tier adds
One-time $29 unlocks the full insight layer, subagent activity inline in the Timeline, lifetime 1.x access and unlimited Macs via Polar checkout.
Where the pricing makes sense
The company stage and team size where Conan's pricing actually pencils out — and where peers do it cheaper.
At $29 once for the Premium unlock (lifetime of 1.x, unlimited Macs), Conan sits well below subscription-priced developer tooling — it roughly matches two months of a typical $15/mo observability add-on. It fits solo developers and small teams already paying for Claude Code who want token and skill visibility without a recurring line item. If you need Windows or Linux coverage, no amount of budget solves it today.
Setup time & first value
How long it actually takes to get something useful out of Conan — broken out by persona, not the marketing-page minute.
Solo developer: download, open, and your existing Claude Code login works immediately — the HUD is live in under five minutes as a free observer. Prompt engineer tuning skills: twenty to thirty minutes to learn the Timeline, Skills & MCP and Context panels well enough to act on them. Team lead auditing spend: plan on an hour for a first real session and a walkthrough of the Context breakdown
Switching to or from Conan
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From raw terminal Claude Code: install Conan, keep your normal Claude login, and let it observe the same terminal rather than switching tools.
- →From ad-hoc context counting: enable auto context tracking in Settings so the Context panel measures the window without manual probes.
- →From separate notes on skill usage: let the Skills & MCP inventory's last-fired timestamps replace your manual log.
- ↗To raw terminal Claude Code: quit Conan and run the CLI as before — nothing about your Claude Code setup is Conan-specific.
- ↗To a non-macOS observability setup: export your PRD.md and CLAUDE.md from Conan Setup and reuse them, since the spec artifacts are plain files.
Integrations
Resources & Guides
Tutorials & Learning
YouTube returned 6 videos for “Conan”, and we withheld 6: 6 could not be judged, because “Conan” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Conan.
Official links
Tools that pair well with Conan
Common stack mates teams adopt alongside Conan, with the specific reason each pairing earns its keep.
Bito
Bito Governor is an AI model router with a code context engine that cuts coding-agent spend by 40-70%.
Honeycomb Query Assistant
Honeycomb Query Assistant turns plain-English questions into HQL against live production telemetry.
Moderne
Moderne applies deterministic recipes for estate-wide code modernization across thousands of repositories.
Featured Head-to-Head Comparisons
Conan vs Spider Cloud
Conan and Spider Cloud serve entirely different needs. Choose Conan if you're a Claude Code power user on macOS who demands real-time observability into every prompt, tool call, and token. Choose Spider Cloud if you need a fast, scalable web scraping API with AI extraction and Browser AI commands for feeding web data into your AI agents. There's no direct overlap; the decision hinges on whether your bottleneck is debugging Claude Code or gathering live web content.
Conan vs Temporal Ai
If you're building durable AI agents or multi-step workflows that need fault tolerance and retries, Temporal AI is the clear choice. For developers deeply into Claude Code who want real-time transparency into prompts, tokens, and tool calls, Conan is an invaluable companion. They serve different purposes: one orchestrates robust backend processes, the other debugs an AI coding assistant. Choose based on your primary need.
Conan vs Voyage Ai
Choose Voyage AI if you need high-accuracy, domain-specific embedding models for enterprise RAG – it's purpose-built for finance, legal, and code retrieval with low-dimensional vectors and long-context support. Choose Conan if you're a Claude Code power user on macOS who demands real-time visibility into token usage, context windows, and skill execution. These tools address completely different problems: Voyage AI powers search and retrieval at scale, while Conan gives you a developer debugging overlay for AI coding sessions.
Alternatives to Conan
View all
Bito
Bito Governor is an AI model router with a code context engine that cuts coding-agent spend by 40-70%.
Honeycomb Query Assistant
Honeycomb Query Assistant turns plain-English questions into HQL against live production telemetry.
Frequently Asked Questions
Used Conan? Help shape our editorial sentiment research.