Wenbu or Labyrinthos: a first reading or a tarot practice?
Compare lessons, spreads, artwork and journals. See how Labyrinthos’s tarot learning features differ from Wenbu’s four tools and Agent.
Compare the workflow first
Labyrinthos describes card-meaning exercises, varied spreads and a reading journal for people building a tarot practice. Wenbu offers one- and three-card readings alongside BaZi, I Ching and Zi Wei, with a shared journal and conversation space. Wenbu’s beginner guides are an introduction, not a full tarot course.
Public pages checked 2026-09-29. This comparison uses official sources; paid and signed-in features have not been tested.
Scroll sideways to see the full comparison. With a keyboard, focus the table and use the arrow keys.
Dimension
Wenbu
Labyrinthos
Learning
Bilingual starter guides, card keywords and follow-up questions for the Agent
Flashcard exercises, a meanings reference and story-based lessons
Draws
78 cards; one- and three-card spreads; optional reversals
Preset spreads and custom spreads
Artwork
78 original AI illustrations with a closer-view option
Several illustrated decks, including physical decks for sale
Journal and cost
Free browser journal, optional account history, JSON export and published AI allowances
Official site lists free basic readings and 100 saved readings per free account; check current subscription terms for extras
Your own AI assistant
Structured card results through MCP, the CLI or a Skill
The pages reviewed do not describe equivalent interfaces
Making a choice
Explore Labyrinthos for regular card practice, more spreads or a physical deck. Try Wenbu for a simple draw, other traditions or results you can use with your own assistant. This is a comparison of public features, without paid account testing or an assessment of predictive accuracy.
Wenbu vs Labyrinthos: choose for the task you have (with a task-evaluation exercise)
Start with the task, then the tool. Labyrinthos's public learning page lists flashcard-style meaning practice, a reading journal and self-analysis of recorded readings for people building a tarot practice. Wenbu's free-access page lists 78-card tarot draws, a local journal, JSON export and calculation interfaces as currently free without payment or sign-up, plus MCP, CLI or Skill access so structured results can go to your own assistant. Wenbu does not call its starter guides a full tarot course. This edition adds a task-evaluation exercise on top of the existing comparison, uses official public pages only, and does not test paid or signed-in features or assess predictive accuracy.
Write the task, the destination of results and the acceptance check first; compare brands last.
Labyrinthos's public pages cover practice, a journal and analysis of records; Wenbu's public pages cover draws, a journal, export and calculation interfaces.
Paid and signed-in experiences were not tested, and no accuracy ranking is offered; check each vendor's current terms.
Random draws cannot be reproduced from the same question; deterministic charts can be recalculated from the same inputs and conventions.
Task-evaluation exercise: four steps from task to acceptance
01Step 1: write the task
State in one sentence what you need: meaning practice, multi-spread work, long-term records, or results for your own assistant.
02Step 2: set the destination
Will results stay here for review, or be exported and connected elsewhere? If export matters, check whether public pages describe it.
03Step 3: verify the public basis
Record only what official public pages state, with the source noted; a page that omits a feature is not proof it is absent.
04Step 4: try each once and review
Run one real task in each free scope, judge it against column four, then decide whether to look further into paid tiers.
This figure explains the order of the exercise: write the task, set the destination of results, verify against public pages, then run one real action in each free scope and review. It describes a decision process, not a product ranking; when a step lacks a public basis, gather more information before choosing.
The existing comparison already lists public features on both sides, but a feature list does not answer which tool fits this particular job. A steadier approach is to write three lines first: what you need to finish, where the result should go, and what counts as done well. Once those lines exist, many choices narrow on their own. If the acceptance check is reciting three card keywords without prompts, the practice path matters more; if it is walking away with a record you can organise later, export and storage matter more.
This step also prevents treating a missing page as a missing feature. The Labyrinthos pages reviewed here do not describe interfaces equivalent to Wenbu's calculation access, but that only means those pages do not cover the capability; it is not evidence that the vendor lacks it. The exercise keeps attention on your own task instead of settling the question for either side.
The exercise: a four-column worksheet and three reader scenarios
Write the evaluation as four columns: task, destination of results, public-page basis, and acceptance check. Column one states the action in one sentence. Column two says whether results stay here for review or need to be exported or handed to another tool. Column three records only what official public pages actually state, with the source noted. Column four sets an observable finish line, such as reciting three card keywords or holding a JSON file after the session. If a column has no basis, you need more research before choosing, not a faster decision.
Three common scenarios fit this worksheet. Scenario one, systematic meaning practice: the acceptance check is reciting keywords, and Labyrinthos's public page mentions flashcard practice and a reading journal, which sits closer to that path. Scenario two, one draw before deciding whether to ask more: the acceptance check is having a draw result, and Wenbu's free-access page lists 78-card tarot draws and a local journal, a shorter entry point. Scenario three, feeding results into your own assistant: the acceptance check is a structured result, and Wenbu's integration page says MCP returns structured results and calculation conventions for the host model to interpret; the Labyrinthos pages reviewed here do not describe equivalent interfaces, which is not the same as lacking them.
Records, cost and uncertainty: what the public pages say
Wenbu's free-access page states that BaZi charts and the five-element diagram, the twelve Zi Wei palaces, I Ching coin casting, 78-card tarot draws, the local journal, JSON export and the knowledge handbook currently need no payment or sign-up, and that MCP and CLI calculation interfaces are also free. AI interpretation runs on a daily allowance; after it ends, you can still cast, draw, save journal entries and export results. Labyrinthos's public learning page says the built-in journal can record digital readings or physical readings made with cards, then review common cards, suits, numbers and reversal percentages over a period. Paid and signed-in features were not tested here, so paid tiers are not compared.
Recording method shapes the choice. Wenbu's journal is stored in the browser and can be exported as JSON, and AI allowances do not affect saved local records. Labyrinthos's journal lives in the app, which suits reviewing trends in one place. The first makes data easier to carry elsewhere; the second makes accumulation easier in one place. The methodology page also notes that traditional meanings, AI inference and real-world context should be kept apart, and that passing software tests means tested behaviour matches stated conventions, not that fortune-telling predictions are validated.
Randomness and reproducibility: one practical rule
Wenbu's methodology page separates results from interpretation: BaZi and Zi Wei are charted from input data, while tarot and online I Ching casting are drawn randomly by the program, and the page keeps the original result. Tarot draws one or three cards with equal probability from 78 cards, without repeats in the same draw; with reversals on, each card independently has a 50 percent chance of appearing reversed. The three positions are present, pull and next step, meant to organise thinking rather than mark a fixed sequence. The model receives card names, orientation and keywords, not card images.
One practical rule follows: deterministic charts can be recalculated from the same inputs and conventions, while random draws cannot be reproduced from the same question. When comparing two tarot results, first check whether it was the same draw, whether reversals were on, and whether position meanings match; when comparing BaZi or Zi Wei, check date, time, time zone and day-boundary convention first. Putting this rule in column three of the worksheet helps you avoid mistaking a different redraw for a tool error.
Hypothetical example: using the worksheet to prepare a reading-group segment
This is a hypothetical situation used to demonstrate the exercise. It involves no real personal data and is not a computed birth chart. Suppose a reader is leading a 20-minute tarot segment at a reading group, needs to practise three-card meanings first, and wants a record they can take away.
Step 1, fill the task column: practise meanings, prepare a three-card flow, keep a record afterwards.
Step 2, fill the destination column: the record should be exportable so it can become notes later.
Step 3, fill the public-page basis column: Labyrinthos's public learning page mentions flashcard practice and a reading journal; Wenbu's free-access page mentions 78-card tarot draws, a local journal and JSON export.
Step 4, fill the acceptance column: recite three card keywords, and hold a readable record file after the session.
Step 5, try each once: run the keyword practice and save one journal entry in Labyrinthos; draw three cards in Wenbu and export the JSON, then judge which record suits the follow-up.
Step 6, review: compare the record with what actually happened, separating card prompts from the discussion that took place.
The point of this hypothetical example is not that one side wins, but that the worksheet makes the choice checkable: practice and journaling are more concentrated in Labyrinthos's public features, while export and connection to your own tools are clearer in Wenbu's public documentation. Neither paid tier was tested; check each current page for the actual experience.
How is this exercise different from the existing comparison?
The existing comparison lists public features on both sides; this exercise puts those features back into your task through a four-column worksheet covering task, destination, public-page basis and acceptance check. It adds no product verdict, only a repeatable decision step that avoids comparing brands before defining the need.
Why can't a random draw be reproduced from the same question?
Wenbu's methodology page says tarot and online I Ching casting are drawn randomly by the program, with the original result kept. Tarot draws one or three cards with equal probability from 78 cards, without repeats in the same draw. Randomness does not guarantee the same card for the same question, so reproducibility means keeping and checking the original result, not redrawing until it matches.
If Labyrinthos doesn't describe an interface, does that mean it has none?
No. The pages reviewed here do not describe equivalent interfaces, which only means those pages do not cover the capability; it is not evidence that the product lacks it. If an interface matters to your workflow, check the vendor's current official documentation and note the result in column three of the worksheet.
Gather more information before choosing. Check official public pages for whether draws, journals, export or interfaces are described; if you still find nothing, leave the column blank and mark it unconfirmed rather than filling it with a guess. A blank is itself a useful finding: it shows the requirement currently lacks checkable information.
A four-column record: task, destination of results, public-page basis and acceptance check. It turns feature comparison into checkable decision steps without settling the question for either side.
A result returned in fixed fields, such as card name, orientation and keywords, so a program or your own assistant can process it rather than reading only prose.
A card presented upside down. Wenbu's methodology page states that with reversals on, each card independently has a 50 percent chance of appearing reversed.
This AI-assisted Wenbu article passed automated structure checks and a separate editorial model pass. It is not independent expert review or a paid-account product test.
Wenbu · daily card ↗Current first-party product UI; not a personal forecast · 2026-10-16 · Fetched text
Run the four-column worksheet once in each free scope
Write task, destination, public-page basis and acceptance check as four lines, complete one real action in each free scope, then review against the acceptance check. That tells you more than comparing brands first.