NEKOVA - AI Native Programming Language

Syntax highlighting, two branded color themes, snippets, and commands for the NEKOVA AI-Native Programming Language.

Built by SYNEKCOT Tech 🇳🇬

Features

• Syntax highlighting for all NEKOVA keywords — think, remember, recall, route, serve, connect, object, class, match, yield, error, shape, every, test, expect, imagine, speak, listen, watch, sandbox, prompt, retry, fallback, observe, mock, const, enum, converse, and more — plus operators added in Phase 24: optional chaining (?.) and spread (...), correctly distinguished from the range operator (..), and Phase 25's using/budget soft keywords inside a think clause
• A real language server (Phase 26) — nekova lsp now backs this extension, replacing what was previously syntax-highlighting-only support: live inline errors as you type (from the real lexer/parser, not a separate approximation), hover docs that resolve to a task/class's actual signature and docstring when it's user-defined, and autocomplete for keywords, builtins, and everything declared in the open file — including type-aware method suggestions right after obj.. Toggle via the nekova.enableLanguageServer setting
• observe blocks and mock think highlighted and auto-indented correctly, plus the |> pipe operator
• Two color themes — NEKOVA Dark and NEKOVA Light — both built from NEKOVA's brand palette of green and white only. AI-native keywords (think, speak, listen, imagine, watch, every) are highlighted in the brightest green in both themes, since that's the language's whole identity. True red/amber are reserved exclusively for actual errors and warnings, so diagnostics stay visually distinct from ordinary syntax in either theme. NEKOVA Light uses a genuine pure-white (#FFFFFF) background, not a lightened grey. Select either via Preferences: Color Theme → NEKOVA Dark / NEKOVA Light
• Branded .nk file icon — a simplified "nk" mark on a small dark backing, designed to stay legible at actual file-icon size (16px), not just at logo size
• prompt blocks highlighted as a declaration (like task/class) only when genuinely defining a prompt — prompt name(...): — while still working correctly as an ordinary variable name elsewhere, matching NEKOVA's own soft-keyword design
• 82 snippets — type think, task, route, obj, class, match, pipeline, parallel, sandbox, taskgen, decorator, shape, every, test, prompt, retry, observe, mock, pipe, const, enum, lettuple, letlist, letdict, optchain, setlit, callkw, thinkerr, taskdoc, converse, thinkshape, thinkbudget, thinkmodel, imaginefile, aiusage, expectsnap, and more
• Run file — press F5 to run the active .nk file
• Watch mode — press Ctrl+F5 to run with --watch (auto-reruns on save)
• Format file — press Shift+Alt+F to run nekova fmt
• Lint file — right-click → NEKOVA: Check File
• Test file — right-click → NEKOVA: Test File (runs the active file, including any test/expect blocks it contains)
• Debug file — right-click → NEKOVA: Debug File
• New project — Command Palette → NEKOVA: New Project (choose template)
• Open REPL — Command Palette → NEKOVA: Open REPL
• Status bar — shows active AI provider, click to run

Requirements

NEKOVA must be installed:

pip install nekova-lang

Commands

Snippets

Settings

NEKOVA Language

# AI is just syntax
think "What should I build today?" as text

# Speak and listen — built in
speak "Hello, world!"
let answer = listen "What city are you in?"

# Named, composable prompts
prompt summarize(text, style="professional", max_sentences=3):
    """
    Summarize the following in a {style} tone.
    Use at most {max_sentences} sentences.
    Text: {text}
    """

let summary = think summarize(article, style="casual")

# Resilience for AI and network calls
retry 3 times with exponential backoff:
    let result = think "analyse this" as json
fallback:
    let result = {error: "unavailable"}

# Tag and trace a block of execution
observe "pipeline run" with tags {user: user_id}:
    let summary = think summarize(document)

# Deterministic AI output in tests
test "classifier":
    mock think as "sports"
    expect classify(text) == "sports"

# Pipe operator — chain transformations left to right
let result = data |> parse() |> filter() |> sort() |> take(10)

# Classes with inheritance
class Animal:
    name: str
    init(name: str):
        self.name = name
    func speak():
        return self.name + " says hello"

let a = new Animal("Lion")
show a.speak()

# Generators
task count(n: int):
    let i = 0
    while i < n:
        yield i
        let i = i + 1

for x in count(5):
    show x

# Built-in test runner
test "math works":
    expect 1 + 1 == 2

# Snapshot testing for AI outputs (Phase 26)
test "greeting shape":
    expect_snapshot(think "Say hi" as text, "greeting")

# Sandbox — safe execution
sandbox strict:
    show 2 + 2
show sandbox_result["safe"]

# Standard library written in NEKOVA
use math
show clamp(15, 0, 10)
show factorial(10)

# Pattern matching with ranges
match status:
    when "ok":    show "All good"
    when "error": show "Something failed"
    else:         show "Unknown"

Release Notes

1.12.0

• Version bump for consistency with NEKOVA core v1.12.0 (Phase 26b, Education Layer). No editor-facing changes in this release — the new tools (nekova learn, nekova explain, nekova translate, nekova classroom, nekova help) are CLI-only. nekova check's two new warnings (comparing to a boolean literal, float equality) surface through the extension's existing Check File command, same as any other checker warning — they are not part of the live LSP diagnostics-as-you-type, which cover syntax errors only
• Welcome message and status bar now report v1.12.0

1.11.0

• Real language server support — this extension now spawns nekova lsp (a hand-rolled JSON-RPC-over-stdio server, part of NEKOVA core as of Phase 26) as a subprocess and wires it up for .nk files: live inline errors from the actual lexer/parser instead of nothing between manual Check File runs, hover docs that resolve to a task/class's real signature and docstring when it's user-defined, and autocomplete for keywords, builtins, and everything declared in the open document — including type-aware method suggestions right after obj.. New nekova.enableLanguageServer setting (default on) to turn it off if needed
• 1 new snippet: expectsnap (expect_snapshot, Phase 26's snapshot testing for AI outputs) — no new syntax-highlighting patterns were needed for it, since it's an ordinary function call and already matched by the extension's existing generic identifier( highlighting rule, same as any other builtin
• Welcome message and status bar now report v1.11.0: real inline errors/hover/autocomplete, nekova fmt --diff, the interactive nekova new wizard, nekova.lock, --why, and snapshot testing

1.10.0

• Syntax highlighting for Phase 25's converse keyword, plus scoped highlighting for the using and budget soft keywords inside a think clause (think "..." using "claude-sonnet", think "..." with budget: 500) — matched the same lookahead-based approach already used for prompt, so ordinary variables named using or budget elsewhere in a program aren't affected
• 6 new snippets: converse, thinkshape (think as a defined shape), thinkbudget, thinkmodel, imaginefile, aiusage
• Welcome message and status bar now report v1.10.0: the docs site going live, plus think ... as <Shape>, cost/token tracking, converse: blocks, and the sandbox prompt-injection guard

1.9.9

• Syntax highlighting for all of Phase 23/24's new syntax: const, enum, optional chaining (?.), and spread (...) — previously ... inside a list/dict would highlight as two range operators (..) plus a stray dot, since spread didn't exist as its own pattern; the range pattern is now correctly checked after spread so the three dots aren't split apart
• 10 new snippets: const, enum, letlist/lettuple/letdict (destructuring), optchain, setlit, callkw (keyword arguments), thinkerr (think with inline error handling), taskdoc (task with a docstring)
• Welcome message and status bar now report v1.9.9, describing what actually shipped: destructuring, keyword arguments, const, spread syntax, optional chaining, enums, and a new Set type, plus more accurate error messages (indentation depth, builtin exceptions) — no extension-side logic changes were needed for the error-message improvements, since those are core-interpreter fixes, but the version and in-editor messaging now reflect them

1.9.8

• Welcome message and status bar now report v1.9.8, announcing Phase 23's correctness fixes (accurate recursion errors, labeled mock AI responses, type-mismatch errors, near-miss variable suggestions) — no extension-side code changes required for these, since they're core-interpreter fixes, but the version and in-editor messaging now reflect them
• Fixed this file — the 1.9.7 entry below was previously garbled (an unclosed bold marker and a stray quote, with no real content) and the Features section's theme description was stale, still describing a single grey-inclusive theme after 1.9.7 had already split it into two grey-free themes

1.9.7

• Added the NEKOVA Light color theme — a genuine second theme, not a toggle: pure white (#FFFFFF) background, dark green-black body text (the one necessary compromise, since literal white text is invisible on white), saturated greens for keywords and accents
• Revised NEKOVA Dark — removed every grey token color (comments, operators, secondary keyword groups) in favor of white and green only. True red/amber remain reserved exclusively for error/warning UI in both themes
• Replaced the .nk file icon with the current simplified "nk" mark — an earlier direct trace of the full NEKOVA logo tested illegible at actual 16px file-icon size, even though it read fine at logo size, so it was rebuilt from simple strokes instead of fine detail like thin rings

1.9.6

• Added observe "label" with tags {...}: highlighting and correct auto-indent (block-opening, like sandbox/retry)
• Added mock think as "..." highlighting — a single-line statement, not a block, so it's intentionally absent from the indent rules
• Added the \|> pipe operator
• Added observe, mock, and pipe snippets (65 snippets total, up from 62)
• Fixed language-configuration.json's onEnterRules — it's a separate list from indentationRules for a different VS Code mechanism (Enter-key behavior vs. paste/reindent), and had silently fallen out of sync since 1.9.5: missing finally, memory, retry, fallback, every, watch, and prompt even though indentationRules already had them

1.9.5

• Added the NEKOVA Dark color theme — pepper-red and gold throughout, with AI-native keywords highlighted brightest
• Added prompt block highlighting as a declaration keyword, but only when genuinely defining a prompt (prompt name(...):) — everywhere else prompt still highlights as a plain identifier, matching its status as a soft keyword at the language level
• Added NEKOVA: Test File and NEKOVA: Debug File commands
• Added prompt, retry, and retrys snippets
• Added retry, fallback, every, watch, and prompt to auto-indent rules after a trailing :; added fallback to the auto-dedent rules alongside catch/finally, since it's a sibling clause of retry
• Fixed extension packaging — the file icon and color theme referenced paths (icons/nk-file.svg, themes/nekova-color-theme.json) that didn't match where those files actually lived in the packaged extension. Fixed and verified by unzipping the built .vsix directly

1.9.4

• Fixed the file-icon theme (nekova-icons) — .nk files were falling back to the generic text-file icon because the theme pointed at an icons/ folder that didn't exist in the packaged extension. The custom .nk file icon now actually renders in the file explorer

1.9.2

• Grammar updated for Phases 15–19: yield, class, error, shape, every, test/expect, imagine, speak/listen, watch, sandbox/strict/relaxed
• Decorator syntax (@) and return type hints (->) highlighted
• Floor division (//) and range operator (..) added to operator highlighting
• Hex (0xFF), scientific (1e5), and underscore (1_000) number literals highlighted correctly
• 24 new snippets — generators, decorators, error types, typed tasks, classes, sandbox blocks, stdlib imports
• 59 total snippets (up from 35)

1.3.0

• Added nekova.runWatch command (Ctrl+F5) — --watch mode
• Added nekova.fmtFile command (Shift+Alt+F) — format file
• Added nekova.checkFile command — static analysis
• Added nekova.newProject command — template picker (default/web/ai/fullstack)
• 35 snippets covering all NEKOVA features
• Grammar updated: object/init/self/extends/super, match/when, route/serve, connect/query, remember/recall/forget, as text|json|list|bool|schema, f-strings, async/await
• Status bar now shows active AI provider
• Welcome message on first install
• nekova.toml activates the extension

1.1.0

• Initial release with syntax highlighting and Run/REPL commands

GitHub · PyPI · License · Built by SYNEKCOT Tech

Licensed under the Business Source License 1.1 — free for personal use, learning, internal tools, and any product written in NEKOVA, with no revenue cap. See the Licensing FAQ for details.