// AI-native development platform

Your entire dev
workflow, autonomous.

Ship features, fix bugs, triage issues, write tests, and generate documentation — all autonomously, inside your existing stack.

Request access See capabilities ↓
feature-runner
koderift run --feature "booking-system"
Generating spec from BA interview…
AI implementing in Docker…
✓ MR opened — 38 files, 94 tests passing
bug-fixer
koderift fix --sentry PROJ-1234
Pulling stack trace, investigating…
Writing targeted fix + regression test…
✓ Fix MR ready — includes regression test
triage-engine
koderift triage --batch
Analysing 12 queued bugs…
Assessing severity, root cause, files…
✓ 8 auto-fixable, 3 need dev, 1 wont-fix
test-writer
koderift tests --check-mr !428
Scanning MR for missing test coverage…
Writing 6 tests for 3 untested methods…
✓ Coverage 74% → 91%, tests MR opened
doc-generator
koderift docs --generate api
Scanning codebase structure…
Generating API reference…
✓ 24 endpoints documented, versioned
mr-guard
koderift guard --mr !512
Reviewing diff, checking test coverage…
Flagging N+1 query in app/models/user.rb…
✓ Review posted, 2 issues flagged
project-intelligence
koderift ask --project payments
"Which controller handles refunds?"
Scanning codebase, building context…
✓ RefundsController#create — 3 files
security-updater Coming soon
koderift secure --auto-patch
CVE-2026-1892 found in nokogiri 1.15.4…
Updating dependency, running test suite…
✓ Patched to 1.15.6, MR opened
10×
Faster than outsourcing
7
Autonomous workflows
100%
Your codebase, your keys
Parallel streams
// capabilities

Seven autonomous workflows.
Zero handoffs.

Feature pipelines

BA interview → spec → AI coding in Docker → GitLab MR. The full cycle from business language to reviewed, deployable code.

Core workflow
Bug fixing & triage

AI triages queued bugs in batch, assesses severity and root cause, then fixes auto-fixable issues autonomously — each with a regression-tested MR.

Sentry integration Auto-route & fix
Test coverage

Checks new MRs for missing tests and writes them automatically. Backfills historic gaps to keep coverage high across your entire project.

MR review Backfill
MR Guard

AI reviews every merge request before it merges — catching missing tests, N+1 queries, unbounded loads, and code smells before your team does.

Code review Auto on MR
Documentation generation

AI reads your codebase and produces user guides, developer docs, and API references — kept in sync as code evolves.

User guides Dev docs API refs
Project Intelligence

Ask questions about any project in plain English. Koderift reasons over your live codebase to answer architecture, ownership, and implementation questions instantly.

Codebase Q&A Instant answers
Security updates Coming soon

Automatically patches vulnerable dependencies when CVEs are published. Updates, tests, and opens an MR — before you even see the advisory.

CVE detection Auto-patch
// features

From business idea
to production MR.

💬
BA Interview
📋
Spec Gen
🐳
Docker Runner
⚙️
AI Implements
Tests Pass
🚀
MR → Staging
// 01
💬
AI-powered BA interview

Business analysts describe the feature in plain language. Koderift conducts a structured AI interview, probing for edge cases, acceptance criteria, and constraints.

Natural language → requirements
// 02
📋
Technical spec generation

The platform fetches your live codebase context on demand. A precise technical spec is generated and routed to dev for sign-off.

File-on-demand context
// 03
⚙️
Autonomous implementation

Koderift's AI agent runs in an ephemeral Docker container — isolated network, matching runtime, private gems resolved. It writes, tests, and refines without babysitting.

Isolated Docker runner
// 04
🚀
Review & deploy

An MR lands in GitLab with a full implementation summary and test results. Developers review, approve, and merge — no onboarding, no context loss.

GitLab MR → staging → prod
// bug-fixing & triage

Triaged, investigated,
and fixed autonomously.

📥
Bug Queue
🤖
AI Triage
🔍
Investigate
🐳
Docker Fix
🔧
Fix MR
// 01
📥
Bugs queue up automatically

Bugs flow in from Sentry webhooks and manual reports. Koderift pulls the stack trace, error context, and affected endpoints automatically — with source-aware context for each.

Sentry webhook Manual reports
// 02
🤖
AI triages in batch

The AI evaluates severity, infers root cause, identifies affected files, and routes each bug: auto-fixable issues go straight to the fix pipeline; complex bugs get flagged for a developer interview.

Root cause analysis Smart routing
// 03
🐳
Autonomous fix in Docker

For auto-fixable bugs, Koderift branches from staging, writes a targeted fix in an isolated Docker environment, and generates regression tests to prevent recurrence.

Isolated environment Regression tests
// 04
🔧
Fix MR ready to review

A merge request lands in GitLab with the fix, regression test, and a full explanation. Your team reviews the diff — the fix is ready before they finish coffee.

Auto MR Regression tests included
// test-coverage

Every MR tested.
Every gap backfilled.

🔀
New MR
🔍
Coverage Scan
🧪
Write Tests
Tests Pass
📈
Coverage Up
// 01
🔍
MR coverage check

When a new merge request lands, Koderift scans the changed files for missing test coverage — untested methods, uncovered branches, and new code paths without assertions.

Automatic on every MR
// 02
🧪
AI writes the tests

The AI agent understands your testing patterns, fixtures, and conventions. It writes targeted tests that match your project's style — not generic boilerplate, real coverage.

Matches your test style Real assertions
// 03
📈
Backfill historic gaps

Beyond new MRs, Koderift can scan your entire project for untested code and backfill missing tests in batches — fixing years of coverage debt incrementally.

Project-wide backfill Incremental batches
// mr-guard

Every MR reviewed.
Before it merges.

🔀
MR Opened
🔍
Diff Analysis
🧠
AI Review
💬
Feedback Posted
Safe to Merge
// 01
🔀
Triggered on every MR

MR Guard fires automatically whenever a merge request is opened or updated in GitLab. No configuration per MR — it watches the whole project.

Automatic trigger GitLab integration
// 02
🧠
AI reviews the diff

The AI reads the full diff alongside the original feature spec. It flags missing tests, N+1 queries, unbounded database loads, security concerns, and logic errors — with the original intent as context.

Spec-aware review Performance checks
// 03
💬
Feedback posted to GitLab

Review comments land directly on the MR in GitLab. Blockers keep the MR in draft; on a clean pass MR Guard approves and undrafts — ready for a human reviewer.

Inline MR comments Auto-undraft on pass
// 04
Nothing slips through

Every MR gets the same rigorous review regardless of who wrote it. Human reviewers spend time on architecture and business logic — not spotting missing indexes.

Consistent quality bar
// project-intelligence

Ask anything about
your codebase.

💬
Question
🔎
Codebase Scan
🧠
AI Reasons
Answer + Files
// 01
💬
Ask in plain English

Type any question about a project — architecture, ownership, implementation details, or where a given piece of behaviour lives. No query language, no setup. Just ask.

Natural language Any project
// 02
🧠
AI reasons over your code

Koderift fetches live codebase context on demand — reading relevant files, tracing call paths, and reasoning across models, controllers, services, and specs to build a precise answer.

Live codebase context Cross-file reasoning
// 03
Answer with source files

You get a clear, cited answer pointing to the exact files and line ranges. Onboard new developers in minutes, answer stakeholder questions instantly, and debug faster.

File citations Instant onboarding
// documentation

Docs that write themselves
and stay current.

🔎
Codebase Scan
🧠
AI Analysis
📝
Doc Gen
💬
Refine
📚
Versioned
// 01
🔎
AI scans your codebase

Koderift analyses your repository structure, models, controllers, services, and API endpoints to understand what needs documenting and how it all connects.

Automatic discovery
// 02
📝
Three doc types generated

User guides for end users, developer documentation for your team, and complete API references — each tailored to its audience with the right level of detail.

User guides Dev docs API refs
// 03
💬
Interactive refinement

An AI-guided interview lets you refine tone, scope, and focus areas. Upload screenshots, add context, and shape the documentation to match your team's standards.

Guided interview
// 04
📚
Versioned & synced

Documentation is versioned with your codebase. When code changes, Koderift detects drift and regenerates affected sections — your docs never go stale.

Auto-sync on code changes
// security  Coming soon

Vulnerabilities patched
before you read the advisory.

🛡️
CVE Published
🔍
Dependency Scan
📦
Update Dep
🧪
Tests Pass
🔒
Secure MR
// 01
🛡️
CVE detected in your deps

Koderift monitors your project dependencies against published CVEs — whether they're Ruby gems, npm packages, or any other package manager. When a vulnerability is found, it identifies the affected package, severity, and the safe version to upgrade to.

Continuous monitoring Severity assessment
// 02
📦
Automatic dependency update

The vulnerable package is updated to the patched version in an isolated environment. Koderift resolves dependency conflicts, updates the lockfile, and verifies compatibility with your stack.

Dependency resolution Lockfile update
// 03
🔒
Tested & ready to merge

Your full test suite runs against the updated dependency. If everything passes, an MR is opened with the CVE details, upgrade diff, and test results — ready for a quick review and merge.

Full test suite CVE details in MR

Ready to automate your entire dev workflow?

Features, bug fixes, MR reviews, test coverage, project intelligence, and documentation — all autonomous, all inside your stack.