What's new in Dolfin 🐬

🤖 Claude can now build with you, not just read

Big step for the Agrafe MCP server 🎉 Connected Claude can now work on your projects, from first file to live deployment.

• Create projects, read and write files, edit package.dlf, and change several files in one go. Claude checks its draft with the same checker as your editor before saving, sees your Problems panel and can apply its quick fixes.
• Deploy and update a deployment, manage its API tokens, and write data into it after a dry run.
• You stay in control: write access is off until you tick it, right by right, on the consent screen when you connect. Every change Claude saves shows in History with an mcp badge, and a new token reaches you through a one-time link instead of the chat.
• Turn on Live view in a project's settings and Claude can query and validate your current files without deploying first.
• The Dolfin and Agrafe guides are built in, so Claude reads the docs of the version you're using.

Read the MCP guide

📤 Turn your queries into spreadsheets

New export declaration: take a query, name your columns, and Dolfin writes the file for you, as CSV, Excel or JSON. Several exports can share one workbook, one sheet each. Report cards, attendance lists, rosters: straight out of your model.

Once deployed, the Deployments tab lists every file with a Download button, also available through the API with your token. Locally, radeau export --out reports/ builds them all.

🔭 Peek inside imported packages

Go to definition on a name from an imported package now opens its source, exactly as in the release you locked, in a read-only tab.

✅ Check your data against your model

• Concepts can now say what valid data looks like: bounds on a field (has age: one int [ >= 0 ]), cross-field check conditions, computed fields with now(), and set of / list of fields. Dolfin turns them into standard SHACL shapes.
• A new Validation tab in Agrafe checks your current package live, or any deployment, and every result is clickable. Deployments show what validation covers.
• Facts in your own files are checked after each save: problems land in the Problems panel and as squiggles on the faulty fact.
• Rules can count: count ?x as ?n then ?n [ > 5 ], like in queries.

📏 Quantities, the standard way

Quantity types are now built in and lowercase: mass, length, duration… The generated Turtle uses QUDT units, so other tools understand your values out of the box. Still declaring concept Mass? The editor spots it and fixes it in one click.

📥 More files Agrafe understands

• GPS tracks: .gpx files can be mapped like any XML file.
• N3 rules: drop a .n3 file and its rules become Dolfin rules, in Agrafe or with radeau init n3.
• Turtle files with no ontology header now import too.

📖 Eqlat has its own guide

Eqlat's documentation is online, linked from every menu bar and the sign-in page.

🔴 See your data come alive

• New Instances tab: pick a few concepts and see their instances as circles, linked by labelled arrows. Drag them around and the picture settles by itself.
• The editor counts for you: · 12 instances above a concept, · 40 uses above a property, rows above a query and a rule.

📚 Richer glossaries

Glossaries now show your queries, what each rule does, a description for each variant and readable cardinalities (exactly one, at least two). In Pangloss, field names in a concept's table link to their property page.

📑 Read your data files at a glance

• CSV files open as a table with a search box that highlights matches and counts rows. Click Edit to change cells directly; your changes save and sync like normal typing.
• JSON and XML files open as a tree you can fold and unfold. Searching opens the branches that contain a match, and a typo in the file is shown with its line and column.
• Large files stay quick: a 5 MB spreadsheet opens, searches and edits without waiting.
• Visitors of a public project see the same table and tree views, read only.

🗃️ Move files around with drag and drop

• Drag a file onto a folder to move it there. Hover a closed folder while dragging and it opens for you.
• Renaming now shows the full path: change the folder part to move the file in the same step.
• + New file here opens the folder so you can type the name straight away, and the Graph pane updates right after a rename, move or delete.
• Inside a mapping, the editor menu offers Open in mapping editor to jump to the form for that exact mapping.
• Projects synced to Git keep their folders on both sides.

📚 Documentation with short videos

The Agrafe and Dolfin language guides now include the matching clip from our video tutorials right next to each topic, so you can read it or watch it. Pangloss and Owlograf have their own guide too, linked from their menu bar and sign-in page.

🦉 Owlograf keeps your layout and follows your edits

• Owlograf now remembers where you placed each box, and shares that layout with the Graph pane in Agrafe. Without edit rights, your arrangement is kept in your browser.
• Changes made in Agrafe show up on the map within a few seconds, no reload needed.
• New concepts land in free space instead of on top of another box.
• Jump to definition opens the file in Agrafe at the right line, and Previous / Next walk through your files one by one.
• Before deleting a box, Owlograf lists everything that still refers to it. Escape closes the right-click menu.

✏️ Turn a draw.io diagram into an ontology

Already sketched your model as boxes and arrows in draw.io? radeau init drawio turns the diagram into a ready-to-use Dolfin project: boxes become concepts, arrows become links between them. It works with UML class diagrams, database tables or plain sketches, and asks you a few questions when the drawing is ambiguous. From then on, the Dolfin files are your model.

🌿 Keep your project in Git

Connect an Agrafe project to a repository on GitHub, GitLab or your own server, and Agrafe keeps a copy of it there as you work. Every release you create becomes a tag, so your developers can follow the model with their usual tools and review changes in pull requests. Available on the Pro and Enterprise plans, from the Package pane.

🧠 Say more with your concepts and rules

• Describe a concept precisely: a Customer is a Person with at least one Order, a Vehicle that is not Electric, or declare that two concepts mean the same thing.
• Rules can follow chains of links, like ancestor of meaning “parent of, any number of times”.
• OWL ontologies that use these constructs now come back into Dolfin intact when you import them.

📱 Graphs on tablets, and in your language

• Pinch to zoom, drag with a finger and long-press for menus: Agrafe and Owlograf now work on tablets and phones, iPad included.
• Graph boxes can show the names from your glossary, with a language picker to switch between, say, English and French.

🐢 Several Turtle files, one project

A turtle-origin project can now hold several Turtle sources. Import them together or add one later: each one keeps its own two-way sync with the Dolfin files.

🦉 Meet Owlograf: explore your ontology as a map

A new app in the suite, made for looking around rather than typing. Sign in with your Agrafe account and open any of your projects, or a public one, as a big interactive graph.

• A magnifying lens lets you zoom on one corner of a large model without losing the overall picture.
• Right-click a box to add, rename or remove concepts and properties right from the graph.
• Optional snap to grid to keep your layout neat.

Open Owlograf

🧩 Build on other people's ontologies

A project can now import another one, yours, a public project or a Turtle ontology on the web, and use its concepts as if they were your own. You pick the project, then the version you want, and it stays on that version until you choose to move. Publish your own stable versions with Create release so others can build on them safely.

🗑️ Faster tidying on the Projects page and the graph

• While dragging a project or folder, a bin appears at the bottom of the page: drop it there to delete it. A folder that still has content asks first; drop it again to delete everything in it.
• A new project or folder scrolls into view, highlighted and ready to name.
• Concepts borrowed from other vocabularies, like foaf:Person, now show up on the graph in an Outside group instead of disappearing.
• Ctrl/Cmd+click on the graph zooms in one step, and the panel divider can be dragged with a finger.

🎬 A new home page and video tutorials

The dolfin.fr home page has been redone to show what Dolfin does in a minute, with real models and a replay of Claude working on one. New to Dolfin? The Agrafe sign-in page now links to our YouTube tutorial series.

📊 Bring your data in by pointing and clicking

Drop a CSV, JSON, XML or Excel/LibreOffice file in your project, right-click it and choose New mapping from file. A form shows your columns next to your concepts, suggests which goes where, and previews the result before you save. Messy data? Mappings can now skip rows, translate codes into proper values and tidy up text on the way in. Claude can also draft a mapping for you through the Agrafe connector.

🔑 One sign-in across the suite

• Pangloss and Eqlat share Agrafe's sign-in page, and bring you back to the page you were on.
• Your projects are listed in the same folders and order as in Agrafe; Eqlat groups your deployments by project.
• Eqlat now tells you when a new version is available, like Pangloss.

💻 Try your model locally with radeau serve

• radeau serve opens the playground in your browser on your own project. Edits in the browser are saved to your files, and changes to the files show up in the page. Add --no-editor for a read-only view.
• radeau check --mapping-dry-run runs every mapping and tells you how many rows were read and skipped, without building anything.
• radeau init now offers the available importers (Turtle, draw.io…) in its menu.

🛶 raft is now radeau

The command-line tool has a new name: radeau. The name raft was already taken on crates.io. Same commands, new name: radeau check, radeau build, radeau fmt. Plugins are now radeau-* executables and read the RADEAU environment variable. Re-run radeau completions to update your shell.

🗂️ A tidier Agrafe workspace

• The editor pane folds away like the sidebar: click ‹ at the start of the tab bar to give the output panel the full width, › to bring it back. Your text, cursor and undo history stay put, and opening a file expands it again.
• New folder now lets you type its name first. Enter creates it, Escape cancels.
• Dolfin files generated from a Turtle import now sit under their source .ttl in the file tree (for imports made from now on).
• Drop a project on a breadcrumb to move it back out of a folder. Dragging a language onto the row just above it now reorders it.
• When sending feedback, drag on the screenshot to send only the part that matters.

🔤 Shorter, correct names in Eqlat results

Eqlat now picks up your deployment's own prefixes, so results read clinic:Patient instead of a long IRI. When namespaces overlap, the most specific one wins, and the full IRI is still one hover away. Public SPARQL visitors get the same prefixes as you.

🕰️ Public projects, every one of them, at any version

• The Public projects page now lists every project made public, not just a selection. Owners without a display name show as "Anonymous", never by email.
• A Version shown picker lets visitors browse a public project as it was at any past deployment: files, graph and references. Works for deployments made from now on.
• Signed in? Public pages show My projects instead of "Sign in", and your projects page links to the public ones. In Pangloss, links in a public glossary stay public, and a View in Agrafe button takes you to the project.

🤝 Invitations, right inside Agrafe

Share a project with someone who already has a Dolfin account and it shows up in their project list, with Accept and Decline buttons. No link to click, and these invitations never expire. You get an email if they decline. People without an account still get the emailed invite link.

Re-sharing with the same address no longer creates duplicate rows.

🕸️ Edit your ontology from the graph

The Graph tab is no longer read-only. Right-click a concept to edit it, rename it (everywhere it's used), add a sub-concept or a property, or delete it. Right-click an arrow to do the same with a property. Every change is written back to your .dlf files, and collaborators see it live.

• Hover an arrow or a field to see its comment, or its Domain → Range signature when it has none.
• New concepts and properties live under a round + button; reset and zoom under a magnifying glass, so the graph keeps its room.
• Reveal in Turtle now works on rules too.

🔧 Queries and exports you can trust

• Queries now produce correct SPARQL for class names (Person, person.Person), for primitive types like [ a string ], and for unions mixing both.
• When two concepts share a field (say name), the Turtle export no longer implies that everything with a name is both. A named vet is not an Animal anymore.
• @timezone Europe/Paris and other named time zones now work, daylight saving included.
• The editor warns when a rule can't be exported as a live N3 rule, and no longer flags a query's answer variables as unused.
• Go-to-definition works in files with spaces or accents in their name, and jumps to the right field when several concepts share one. Renaming a file no longer doubles its references.
• Fixed: the graph could fail to render on some unions.

⚖️ Rules that actually fire, and more units

• Rules with quantifiers (none, all, at least 2…) are now exported as live N3 rules instead of being commented out. When no faithful encoding exists, the rule stays a comment with the reason, and raft build warns you.
• = and != in a rule now work on text, booleans and resources, not just numbers and dates: name [ = "Bob" ] finally matches Bob. between N, M parses too.
• New units: knot, nautical mile, US pint, carat, technical atmosphere, hectare and dalton, plus every SI prefix from quecto to quetta. kt is the knot, not a kilotonne, and units that never take a prefix (like kg or h) now reject one.

🧰 A sturdier raft

• raft fmt no longer loses anything: rules, queries, cardinalities and comments come out exactly as written, and running it twice changes nothing.
• raft check now runs the full package-wide analysis, the same one the editor uses.
• Run raft with something missing and it asks only for what's missing, keeping the options you already typed. Plugins show up in --help and shell completion.

🎯 Dolfin figures out which property you mean

When two concepts each have a name, you no longer need to qualify it. In facts, rules and queries, Dolfin looks at the type of the thing you're describing and picks the right property, in the Turtle and SPARQL it generates too. It only falls back to the old behaviour when it genuinely can't tell.

• A bare fact or enum value declared in another file now resolves to that file, and a value that exists nowhere in the package is flagged.
• The linter offers a one-click rename to camelCase across the whole package, catches rule variables used only once, and unknown concepts inside rule and query bodies.

🧭 Find your way around the Agrafe editor

• Back / forward through where you've been, with Alt+[ / Alt+] or the ‹ › buttons. Each file remembers its caret and scroll position when you switch tabs.
• Go-to-definition and find-references now open the other file's tab when the target lives there. Every declaration shows an "N references" line you can click.
• Viewers get a proper read-only editor, and actions you're not allowed to take (deploy, rename, delete, quick fixes…) are hidden instead of failing.
• Fixed: typing at the end of a file while collaborating could pile up stray blank lines.

🤖 Talk to your ontologies from Claude

Agrafe is now an MCP server. Connect Claude (on claude.ai or Claude Desktop) to your account and it can list your deployments, describe a concept, read a glossary and run SPARQL queries, on your projects and on those shared with you. Built-in prompts help it explore an ontology, write a query or explain a concept.

The connection snippet is in Settings → MCP. Queries are capped at 10 s and 1000 rows.

Read the Agrafe guide

🐢 Start from Turtle, keep it in sync

Import an existing .ttl ontology and Agrafe keeps the original alongside the Dolfin translation. Edit in Dolfin and the Turtle follows along automatically.

• Right-click anything (a concept, a field, an enum variant, a fact) and Reveal in Turtle to see the matching statement, or go the other way with Reveal in Dolfin.
• Re-translate from the original with a preview first, and undo if you change your mind.
• A namespace declared with a trailing / is now kept as is, instead of gaining an extra #.

📖 Agrafe gets its own user guide

Workspace, panels, menus, the Problems list, Turtle-origin projects and MCP: all documented at agrafe.dolfin.fr/docs. The language docs stay at dolfin.fr/docs, and the onboarding tours now link to both.

✉️ Account emails you can count on

• Forgot your password? Reset it from the login page. New accounts get an email to confirm their address.
• A heads-up before your license expires, and when you get close to (or reach) your project or deployment quota.
• A note when someone accepts your invite, or when you're removed from a project.
• Paid users can now request a license renewal from the license page.

📊 Separate room for public projects

Public and private projects now count against separate allowances, and public ones get a much wider one, files included. Making a project public never eats into your private quota. The Usage and License tabs show a bar for each.

🌐 Public projects, browsable from every app

Mark any project public from its settings, and anyone can now open it, no Dolfin account needed.

A public project gets a read-only page in Agrafe (file tree, source, compiled graph, and a SPARQL console), a browsable glossary in Pangloss, and a queryable deployment in Eqlat, all reachable straight from each app's new Public projects page.

🧬 Two language upgrades: unions and multilingual descriptions

• Dolfin now understands union types in domain and range: write (Agent or Post) when a property can point at more than one kind of thing. Imported ontologies that use OWL's unionOf pick this up automatically, and the graph view draws it as a small ∪ node fanning out to each option.
• package.dlf's description can now be written in multiple languages too, just like concept and property descriptions
  description@fr "..." alongside the plain one.

🐫 camelCase for property names

Property names are now linted as camelCase instead of snake_case, matching how OWL and RDF do it. Dolfin properties round-trip through both. This is a warning only, so nothing breaks; rename at your own pace.

📁 Organize your projects into folders

• Organize your projects into folders, with nested subfolders. Drag a project (or a whole folder) to move it, or drop it into a folder's carousel to file it away.
• Sort your list manually, by most recently opened, or by most recently modified, pick from a single dropdown.
• A new search box finds any project by name or description across every folder, showing you exactly where it lives.

🔤 Use any word as a name

Want to name a concept or property date or mapping? Wrap it in backticks
`date`
and Dolfin treats it as a plain name even though it's normally a keyword. The editor offers this as a one-click fix whenever a keyword gets in your way, and autocomplete suggests the escaped spelling too.

🛡️ Safer downloads, full keyboard reach in Agrafe

• Downloading Turtle from a project that fails to compile no longer hands you a broken file, the last successful export stays available instead.
• The file tree is now fully keyboard-navigable: arrow keys move between files and folders, one tab stop for the whole tree.
• Package settings now warns before discarding unsaved edits, and the plan/quota screen is clearer about what you're using and what you're allowed.

🛠️ A big usability pass for Agrafe

Agrafe got a broad usability review, and a long list of fixes:

• Deploying now hands you a ready-to-run curl example with your token already filled in, plus a link straight to the full API reference.
• A new project offers two starting points: load an example or paste Dolfin you already have. No more empty screen with nothing to click.
• The editor now indents automatically after a block opener like concept X:, and error messages carry the fix suggestion, not just the problem.
• Below tablet width, the layout adapts: the side rail moves under the editor and the file tree becomes a drawer, instead of squeezing everything into a column.
• A failed compile no longer blanks the graph or SVG view — your last working result stays visible, dimmed, with a note on what broke.
• New guided tours (skip any time), full keyboard support for every dialog (Escape closes, Tab stays inside), and a password field with a show/hide toggle and a strength meter.

👥 See who's viewing, in Pangloss too

Pangloss glossary pages now show live avatars for everyone currently viewing (the same little faces you already see in Agrafe) and a new Shared with me section lists projects a teammate has invited you to, even if you don't own them.

🖱️ Live cursors and shared undo on the graph view

The graph view (SVG tab) now shows everyone's cursor as they move around, just like the text editor already did.

Undo and redo are now shared too: Ctrl+Z undoes the most recent change to the file, even if a teammate made it, instead of only your own local edits.

📝 Write Markdown alongside your models

Agrafe's editor now understands Markdown: open a .md file anywhere in your project and get proper headings, bold, links, and code formatting instead of Dolfin's syntax rules.

🖼️ Save your graph, keep your layout

• Download your project's graph as an image straight from the output tab.
• Agrafe now remembers your workspace: a collapsed sidebar and your panel sizes stay just the way you left them next time you open a project.

🕰️ Go back to any earlier version

Agrafe now keeps a full history of your project. Every save is remembered, so you can look back at what changed, compare an earlier version side by side with where you are now, and restore any past state in one click, even edits you made together with teammates.

Rapid saves are grouped together so the timeline stays easy to read, and a little branch diagram shows how your versions connect.

✍️ Edit together, in real time

Several people can now write in the same file at the same time and see each other's changes appear instantly, no refreshing and no more overwriting a teammate's edit by accident.

Everyone's cursor shows up with a little name tag, so you always know who's working where.

👀 See who's looking at a project with you

Shared projects now show little avatars for everyone currently looking at them, updated live as people come and go. Open the same project on your phone and your laptop, and it still only counts as you once.

🏷️ Cleaner names for well-known vocabularies

When importing data that uses standard, widely-known vocabularies (like RDF, OWL, SKOS, or FOAF), Dolfin now recognises them and gives them their proper short name instead of an ugly technical alias.

🌍 Describe your model in more than one language

You can now write a concept or property's description in several languages at once, each on its own tagged comment line (like #en> and #fr>). The editor auto-continues the right language tag as you type, and can fold each language block away so long multilingual comments don't clutter your file.

Pangloss glossary pages get a language selector to match: pick a language and every label and definition switches with it, falling back to English for anything not yet translated.

📐 Real units and quantities in your models

You can now attach real-world units to your data: write weight: 72 unit.Kilogram and Dolfin understands it as an actual quantity, not just a number. Compare, convert, and reason about values across compatible units automatically.

Currencies and other units that shouldn't be mixed up (like different counting units) are kept properly separate, so you can't accidentally compare apples to oranges.

The editor now helps you as you type: autocomplete and hover descriptions for common physical units and dimensions, right where you need them.

📥 Bring your existing data in

You can now import a project straight from a Turtle file, so data you already have elsewhere doesn't need to be retyped from scratch.

The project list got a refresh too: your projects show as clean cards, action buttons appear when you hover, shared projects have their own section, and deleting now needs a deliberate hold (no more accidental clicks).

Compiled output can also be downloaded directly from the Playground.

📅 Dates and times that just work

You can now write real dates, times, and durations in your models, the way you'd actually say them. Type date(June 1st 2026), time(2:30 PM), or duration(1y 6mo) and Dolfin turns it into the proper machine format for you.

No more guessing whether 01/06/2026 means June or January: set @locale d/m/y once at the top of a file (and @timezone for your local time), and every date in that file follows suit.

And if you mistype a date, the editor now points to exactly where — instead of quietly saving something wrong.

✏️ A much better editing experience

The project editor got a lot more comfortable:

• A proper file tree, a bit like VS Code: rename a file in place, right-click for actions, drag a whole group of files together, and hold-to-delete so nothing disappears by accident.
• A project-wide search bar in the sidebar. Find any concept, property, rule, fact, or even comment text across all your files, and jump straight to it.
• A Problems panel showing every error, warning, and hint in one place. Click any item to jump to the exact spot.
• One-click fixes. Mistype a name and you'll often see a "did you mean Person?" suggestion with an Apply button that fixes it for you, even across files.
• Live graph view of your project: pan, zoom, drag nodes around, and reset. Your layout is remembered between sessions, and clicking a node jumps to its definition.

⌨️ A nicer code editor everywhere

• Ctrl+F search inside the editor.
• Fold away long comment and annotation blocks.
• Smarter typing: comments and annotations continue automatically on the next line.
• Richer hover tooltips with proper formatting and colour.
• A heads-up banner when a new version is available, so you're never stuck on a stale page.

🚀 Publish and query your data

• Deployments now snapshot your project so a live endpoint is stable and won't shift under you. The screen shows the ID and access token clearly, with copy buttons.
• Each deployment reports how many schema vs. data triples it holds, so you can see your dataset at a glance.
• Deployed endpoints now answer questions using your rules, not just the raw facts, results reflect what your project actually means, not only what was literally typed in.

🔎 Eqlat: a brand-new way to ask questions

We launched Eqlat, a visual query tool built on top of sparnatural. Instead of writing query code, you build questions by clicking through your concepts and their relationships, and see the results immediately. It signs you in with your existing Dolfin account, so there's nothing extra to set up.

Open Eqlat →

💬 Ask questions inside Dolfin itself

The Dolfin language now has its own query syntax. Describe the questions you care about right alongside your model, in plain Dolfin, and the tools turn them into real database queries for you. You can also write relationships "backwards" (e.g. "is actor of") wherever it reads more naturally.

📖 Your glossary, easier to navigate

• Glossary pages are now shareable by link, and the browser back/forward buttons work as you'd expect.
• Click any reference (a parent concept, a related property, a domain or range) to jump straight to it.
• Inherited fields are grouped neatly by where they come from, so it's obvious what a concept gets from its parents.
• Mentions like {Person} inside your descriptions now become live links automatically.
• Rules now appear in the glossary, and every entry has an "Edit in Agrafe" link that takes you to the exact line in the editor.

🎨 One look and feel everywhere

The Playground, Pangloss, Agrafe, and the new query app now share the same design and a single top navigation bar.

A new feedback button is on every app too. It even grabs a screenshot of what you're looking at, so telling us about something odd takes seconds. And you can always choose not to send the screenshot.

🤝 Share your projects with teammates

You can now invite people to your projects as a Viewer or an Editor. Shared projects show up in everyone's list with a clear "Shared by …" label, and you can manage members or leave a shared project from settings.

⚡ A faster, richer Playground

The Playground was rebuilt around a modern editor: work on several files at once, browse your glossary in a tab, and import or export your whole project as a zip, still entirely in the browser, nothing to install.

💡 Error messages that actually help

When something doesn't compile, you now get a clear message pointing at the exact file and line. And instead of a blank page, you still see as much of your glossary as could be built.

📖 Pangloss: your glossary as a web page

Pangloss turns a project into a clean, browsable glossary you can publish and share by link, and revoke any time. Readers get a friendly view of your concepts and properties without ever touching the source.

🪟 Edit and see your graph, side by side

Work on your model and watch its graph update right next to it, with a split you can resize. A project-wide view shows everything at once and gently dims the files you're not focused on.

🔐 Accounts and saved projects

Create an account, log in, and keep your projects in one place with Agrafe. Manage your password and account settings, and pick up right where you left off from any machine.

🧩 Dolfin inside VS Code

A standalone VS Code extension brought Dolfin to a familiar editor, no native install, no fuss. Hover over anything to read its description, and jump between definitions across files.

🗺️ A clearer picture of your model

The diagram view got sharper: concept boxes focus on the essentials, arrows connect concepts across ontologies, and you can filter what's shown to find your way around large models.

🔧 Rename across your whole project

Rename a concept or property once and every mention updates — even across different files. One-click fixes now reach throughout your project, not just the file you're looking at.

📚 Documentation for every check

Every built-in check now comes with clear documentation: what it looks for, how to fix it, and how to write your own.

✅ Catch mistakes as you write

Built-in checks flag common problems and suggest tidier naming, and you can switch individual ones on or off for your team. A new formatter keeps every file neat with a single command.

🐬 Dolfin gets going

The Dolfin modelling language take shape : a simple, readable way to describe your concepts, their properties, and how they connect. And you can try it straight away in a browser Playground, nothing to install. You can also add rules, so the tools work out new facts from what you've already stated.