For Engineering Managers & Team Leads

New Engineers Ship CodeIn Days, Not Months

Stop losing 30% of senior engineer time to onboarding questions. Probe gives new team members instant access to architectural context, patterns, and ownership — from day one.

Self-serve knowledgeLiving documentationZero knowledge lossAny LLM

Free and open source. Business and Enterprise plans available.

What changes for your team
Before
2-3 months to first meaningful contribution
Senior engineers interrupted constantly
Knowledge locked in people's heads
After
First PR in days, not weeks
Self-serve codebase answers
Always-current architecture docs
Built for engineering teams
On-prem Any LLM Open source

Why teams choose Probe for onboarding

Radically Faster Ramp-Up

New engineers ask questions and get accurate answers instantly. No waiting on Slack. No pulling senior engineers into meetings. First meaningful PR in days.

Living Documentation

Architecture diagrams and dependency maps generated from actual code, not stale wikis. Always current. Always accurate. Always accessible.

Zero Knowledge Loss

When engineers leave or rotate teams, their knowledge stays behind. Captured in searchable, queryable AI that understands your codebase.

Reclaim Senior Engineer Time

Stop burning 30% of expert time on "where is this?" and "how does this work?" questions. Let them build instead of being bottlenecks.

The Real Problem

The onboarding tax is killing your velocity.

Every new hire costs you weeks of ramp-up time and months of senior engineer productivity. It's not sustainable.

01

The Ramp-Up Tax

6-8 weeks before a senior engineer is productive. 3+ months for complex systems. They're reading code, asking questions, and making small commits. You're paying full salary for partial output. Scale that across every new hire and team rotation.

02

The Interruption Spiral

Every "quick question" to a senior engineer costs 23 minutes of context-switch time. New hire asks where authentication lives. Senior engineer stops, explains, loses their mental stack. Five questions a day. Two hours gone. Multiply by every team member onboarding.

03

The Knowledge Cliff

Your best engineer knows the billing system inside-out. They leave. Their knowledge leaves with them. The next person spends months rediscovering patterns, gotchas, and design decisions. You lose institutional memory with every departure.

What new engineers actually ask Probe

Real questions from onboarding developers. The difference: they get answers in seconds, not days.

Onboarding
Generate architecture overview for new team member
Your system has 12 core services. Frontend: customer-portal (React), admin-dashboard (Vue). Backend: api-gateway (entry point), auth-service (handles login/SSO), user-service (profile management), billing-service (payments), notification-service (email/SMS), analytics-service. Data stores: PostgreSQL (primary), Redis (caching/sessions), S3 (file storage). Key patterns: event-driven via RabbitMQ, circuit breakers in gateway...
Question
Create service dependency map for the payments domain
Payment flow dependencies: billing-service (core) depends on → payment-gateway (Stripe integration), fraud-detection-service (risk scoring), user-service (account validation), notification-service (receipts). External dependencies: Stripe API, tax-calculation-service. Database: billing_db (PostgreSQL). Events published: payment.completed, payment.failed, refund.processed. Consumed by: analytics-service, reporting-service...
Question
What should I know before touching the auth module?
Auth module is critical path for all API requests. Key things: 1. Uses JWT tokens with 1-hour expiry. 2. Session state in Redis (cluster mode). 3. SSO integration via SAML (enterprise customers). 4. Rate limiting: 100 login attempts/hour per IP. 5. Known issues: PROJ-4201 (refresh token edge case), PROJ-4890 (SSO redirect bug). 6. Testing: always test with both local and SSO flows. 7. Owner: Identity team, contact: @sarah-chen...
Onboarding
How do I set up local development for this service?
Local dev setup for api-gateway: 1. Prerequisites: Docker, Node 18+, PostgreSQL. 2. Clone repo and run `npm install`. 3. Copy .env.example to .env, set DATABASE_URL and REDIS_URL. 4. Run `docker-compose up -d` for dependencies. 5. Run migrations: `npm run migrate`. 6. Seed test data: `npm run seed`. 7. Start dev server: `npm run dev` (runs on :3000). 8. Run tests: `npm test`. 9. Common issues: if Redis fails, check Docker network...
Question
Who owns the billing module and what's its history?
Billing module owned by Payments team. Primary contacts: @james-liu (tech lead), @priya-kumar (senior engineer). Created Q2 2022 (originally part of monolith, extracted in PROJ-2100). Major milestones: v1 (Stripe integration), v2 (multi-currency support, PROJ-3400), v3 (subscription billing, PROJ-4200). Known tech debt: legacy invoice generation (needs refactor, tracked in PROJ-5100), race condition in concurrent payments (workaround in place)...
Question
What are the testing patterns used in this repo?
Testing patterns: 1. Unit tests: Jest, target 80% coverage. 2. Integration tests: test DB interactions with test containers. 3. E2E tests: Playwright for critical user flows. 4. Mocking: use jest.mock() for external APIs, never mock internal modules. 5. Factories: use test/factories for test data (UserFactory, OrderFactory). 6. Database: tests run against isolated test DB, auto-cleanup after each test. 7. CI: all tests must pass before merge...
Onboarding
Generate architecture overview for new team member
Your system has 12 core services. Frontend: customer-portal (React), admin-dashboard (Vue). Backend: api-gateway (entry point), auth-service (handles login/SSO), user-service (profile management), billing-service (payments), notification-service (email/SMS), analytics-service. Data stores: PostgreSQL (primary), Redis (caching/sessions), S3 (file storage). Key patterns: event-driven via RabbitMQ, circuit breakers in gateway...
Question
Create service dependency map for the payments domain
Payment flow dependencies: billing-service (core) depends on → payment-gateway (Stripe integration), fraud-detection-service (risk scoring), user-service (account validation), notification-service (receipts). External dependencies: Stripe API, tax-calculation-service. Database: billing_db (PostgreSQL). Events published: payment.completed, payment.failed, refund.processed. Consumed by: analytics-service, reporting-service...
Question
What should I know before touching the auth module?
Auth module is critical path for all API requests. Key things: 1. Uses JWT tokens with 1-hour expiry. 2. Session state in Redis (cluster mode). 3. SSO integration via SAML (enterprise customers). 4. Rate limiting: 100 login attempts/hour per IP. 5. Known issues: PROJ-4201 (refresh token edge case), PROJ-4890 (SSO redirect bug). 6. Testing: always test with both local and SSO flows. 7. Owner: Identity team, contact: @sarah-chen...
Onboarding
How do I set up local development for this service?
Local dev setup for api-gateway: 1. Prerequisites: Docker, Node 18+, PostgreSQL. 2. Clone repo and run `npm install`. 3. Copy .env.example to .env, set DATABASE_URL and REDIS_URL. 4. Run `docker-compose up -d` for dependencies. 5. Run migrations: `npm run migrate`. 6. Seed test data: `npm run seed`. 7. Start dev server: `npm run dev` (runs on :3000). 8. Run tests: `npm test`. 9. Common issues: if Redis fails, check Docker network...
Question
Who owns the billing module and what's its history?
Billing module owned by Payments team. Primary contacts: @james-liu (tech lead), @priya-kumar (senior engineer). Created Q2 2022 (originally part of monolith, extracted in PROJ-2100). Major milestones: v1 (Stripe integration), v2 (multi-currency support, PROJ-3400), v3 (subscription billing, PROJ-4200). Known tech debt: legacy invoice generation (needs refactor, tracked in PROJ-5100), race condition in concurrent payments (workaround in place)...
Question
What are the testing patterns used in this repo?
Testing patterns: 1. Unit tests: Jest, target 80% coverage. 2. Integration tests: test DB interactions with test containers. 3. E2E tests: Playwright for critical user flows. 4. Mocking: use jest.mock() for external APIs, never mock internal modules. 5. Factories: use test/factories for test data (UserFactory, OrderFactory). 6. Database: tests run against isolated test DB, auto-cleanup after each test. 7. CI: all tests must pass before merge...
Question
Walk me through how a user signup flows through the system
User signup flow: 1. Frontend (signup-form.tsx) POST to /api/auth/register. 2. API gateway validates request, forwards to auth-service. 3. Auth-service checks email uniqueness, hashes password (bcrypt), creates user record in users table. 4. Publishes user.created event to RabbitMQ. 5. User-service consumes event, creates profile record. 6. Notification-service consumes event, sends welcome email. 7. Returns JWT token to frontend. Full flow takes ~200ms...
Question
What conventions does this team follow for error handling?
Error handling conventions: 1. Use custom error classes (BusinessError, ValidationError, SystemError) from lib/errors.ts. 2. Never throw raw strings, always throw typed errors. 3. API responses: consistent format {error: {code, message, details}}. 4. Logging: use logger.error() with structured context (userId, requestId, etc). 5. Client errors (4xx): don't log stack traces. Server errors (5xx): full stack + context. 6. Database errors: wrap in SystemError, never expose raw SQL...
Question
What changed in this service in the last month?
Last month changes in billing-service: PR #892: Added support for Apple Pay (merged 12 days ago). PR #901: Fixed tax calculation rounding bug (merged 8 days ago, deployed to prod). PR #905: Upgraded Stripe SDK to v12 (merged 5 days ago, in staging). PR #910: Refactored invoice generation for performance (in review). Config changes: increased payment timeout from 30s to 60s (CONF-440). Known issues introduced: none. Team velocity: 4 PRs/week average...
Automation
Generate README for a service that has none
Generated README.md for notification-service: Overview: Handles all outbound notifications (email, SMS, push). Tech stack: Node.js, TypeScript, Bull (job queue), SendGrid, Twilio. Setup: npm install, configure .env, run docker-compose, npm run dev. Architecture: REST API + background workers. Key endpoints: POST /send-email, POST /send-sms. Queue: notification-queue (Redis-backed). Workers: email-worker, sms-worker. Testing: npm test. Deployment: k8s manifests in deploy/...
Question
What are the common pitfalls in this codebase?
Common pitfalls: 1. N+1 queries: always use .includes() for eager loading in user-service. 2. Race conditions: billing-service has known concurrency issues, use row-level locks. 3. Memory leaks: notification-service workers don't close connections properly, monitor memory. 4. Auth tokens: don't forget to validate expiry in middleware. 5. Feature flags: always check LaunchDarkly before deploying new features. 6. Database migrations: always run migrations before deploy, never after...
Question
How do feature flags work in this project?
Feature flags via LaunchDarkly. Client initialized in lib/feature-flags.ts. Usage: await featureFlags.isEnabled('new-checkout-flow', userId). Server-side flags: evaluated on backend. Client-side flags: streamed to frontend. Flag naming: kebab-case, prefix with team (payments-new-flow). Environments: dev (all flags on), staging (test rollouts), prod (gradual rollout). Best practice: wrap new features in flags, clean up flags after 100% rollout...
Question
Walk me through how a user signup flows through the system
User signup flow: 1. Frontend (signup-form.tsx) POST to /api/auth/register. 2. API gateway validates request, forwards to auth-service. 3. Auth-service checks email uniqueness, hashes password (bcrypt), creates user record in users table. 4. Publishes user.created event to RabbitMQ. 5. User-service consumes event, creates profile record. 6. Notification-service consumes event, sends welcome email. 7. Returns JWT token to frontend. Full flow takes ~200ms...
Question
What conventions does this team follow for error handling?
Error handling conventions: 1. Use custom error classes (BusinessError, ValidationError, SystemError) from lib/errors.ts. 2. Never throw raw strings, always throw typed errors. 3. API responses: consistent format {error: {code, message, details}}. 4. Logging: use logger.error() with structured context (userId, requestId, etc). 5. Client errors (4xx): don't log stack traces. Server errors (5xx): full stack + context. 6. Database errors: wrap in SystemError, never expose raw SQL...
Question
What changed in this service in the last month?
Last month changes in billing-service: PR #892: Added support for Apple Pay (merged 12 days ago). PR #901: Fixed tax calculation rounding bug (merged 8 days ago, deployed to prod). PR #905: Upgraded Stripe SDK to v12 (merged 5 days ago, in staging). PR #910: Refactored invoice generation for performance (in review). Config changes: increased payment timeout from 30s to 60s (CONF-440). Known issues introduced: none. Team velocity: 4 PRs/week average...
Automation
Generate README for a service that has none
Generated README.md for notification-service: Overview: Handles all outbound notifications (email, SMS, push). Tech stack: Node.js, TypeScript, Bull (job queue), SendGrid, Twilio. Setup: npm install, configure .env, run docker-compose, npm run dev. Architecture: REST API + background workers. Key endpoints: POST /send-email, POST /send-sms. Queue: notification-queue (Redis-backed). Workers: email-worker, sms-worker. Testing: npm test. Deployment: k8s manifests in deploy/...
Question
What are the common pitfalls in this codebase?
Common pitfalls: 1. N+1 queries: always use .includes() for eager loading in user-service. 2. Race conditions: billing-service has known concurrency issues, use row-level locks. 3. Memory leaks: notification-service workers don't close connections properly, monitor memory. 4. Auth tokens: don't forget to validate expiry in middleware. 5. Feature flags: always check LaunchDarkly before deploying new features. 6. Database migrations: always run migrations before deploy, never after...
Question
How do feature flags work in this project?
Feature flags via LaunchDarkly. Client initialized in lib/feature-flags.ts. Usage: await featureFlags.isEnabled('new-checkout-flow', userId). Server-side flags: evaluated on backend. Client-side flags: streamed to frontend. Flag naming: kebab-case, prefix with team (payments-new-flow). Environments: dev (all flags on), staging (test rollouts), prod (gradual rollout). Best practice: wrap new features in flags, clean up flags after 100% rollout...

Three things that transform onboarding

01

Instant Architecture Maps

Auto-generated service dependency diagrams, data flow maps, and ownership directories — always current, never stale.

  • "Show me how the payment flow works across all services"
  • "Which teams own which parts of the authentication system?"
  • "What databases does the billing module depend on?"
  • "Map every service that calls the user API"

New engineers get a mental model of the entire system in hours, not weeks. They understand the architecture before writing their first line of code. Probe reads the codebase semantically — tracking function calls, API endpoints, database queries, and message queues — to build a living map of your system.

Service dependency graphsVisual maps of how services connect and communicate
Ownership directoryWho owns what, contact info, team responsibilities
Data flow analysisTrace how data moves through your entire system
02

Self-Serve Knowledge Base

New engineers ask questions about the codebase and get answers instantly — no Slack messages, no waiting, no interrupting teammates.

Stop the interruption tax. Instead of asking "where does authentication live?" in Slack and waiting for a response, new hires query Probe and get an answer in seconds. The knowledge base understands your code semantically, pulling context from implementation, tests, commits, and tickets.

Senior engineers reclaim hours every week. New engineers get unblocked instantly. Knowledge flows freely without bottlenecks.

Code-aware answersResponses grounded in actual implementation, not documentation that might be outdated
Pattern recognitionLearn team conventions for testing, error handling, API design, and more
Historical contextUnderstand why decisions were made, what alternatives were considered
03

Living Documentation

Documentation that stays current because it's generated from actual code, not written by hand and forgotten.

Wikis go stale. READMEs drift. Architecture diagrams become lies. Living documentation is always accurate because it's derived directly from the code. When the code changes, the documentation updates automatically.

Generate architecture overviews, setup guides, API documentation, and onboarding runbooks on-demand. Every new hire gets the latest, most accurate picture of the system.

Auto-generated guidesSetup instructions, architecture overviews, deployment runbooks
Always currentDocumentation reflects actual code, not wishful thinking
Role-specific viewsFrontend engineer sees frontend context, backend engineer sees backend

Workflow packs for every onboarding scenario

Pre-built automation workflows you can deploy immediately. Customize for your team. Version like code. Improve over time.

Day 1

Day 1 Onboarding

New hire joins the team. Automatically generate a personalized getting-started guide for every repository they'll touch. Architecture overview, local setup, key patterns, who to ask for help.

  • Architecture overview document
  • Local development setup guide
  • Team conventions and patterns
  • Ownership and contact directory
Week 1-2

New Project Ramp-Up

Engineer assigned to unfamiliar codebase. Auto-generate deep-dive documentation: service architecture, dependency map, data models, testing patterns, deployment process, and known gotchas.

  • Service-specific architecture doc
  • Dependency and integration map
  • Key patterns and conventions
  • Common pitfalls and workarounds
Team Changes

Team Transfer

Engineer rotating to new team or departing. Automatically capture their knowledge: what they built, why decisions were made, where technical debt lives, what needs attention next.

  • Knowledge transfer document
  • Technical debt inventory
  • Ownership handoff checklist
  • Historical decision context
Continuous

Knowledge Audit

Scan your entire codebase to detect undocumented critical systems, single points of knowledge failure, and areas where one engineer leaving would cause major knowledge loss.

  • Undocumented critical systems
  • Knowledge concentration analysis
  • Documentation gap report
  • Risk mitigation recommendations

Built for engineering teams at scale

On-Premises Deployment

Runs entirely inside your infrastructure. Code never leaves your environment. Full data sovereignty and compliance compatibility.

Any LLM Provider

Use your preferred model — Claude, GPT, open-source, or self-hosted. No vendor lock-in. Switch providers without changing workflows.

Full Audit Trail

OpenTelemetry instrumentation captures every query, every workflow run, every decision. Complete traceability for compliance and debugging.

Open Source Core

The core engine is open source and auditable. You can inspect exactly how your code is being processed. No black boxes.

Open Source vs Enterprise

Start with the open-source core to evaluate the technology, then scale to enterprise when you need multi-repo capabilities.

Probe Open Source

Free forever

The core code intelligence engine. Perfect for individual developers or small teams exploring a single codebase.

  • Single-repository understanding — Ask questions about one repo at a time
  • Semantic code search — Understands code as code (functions, classes, dependencies), not just text
  • No indexing required — Works instantly on any codebase, runs locally
  • MCP integration — Use with Claude Code, Cursor, or any MCP-compatible tool
  • Any LLM provider — Claude, GPT, open-source models — your choice
  • Privacy-first — Everything runs locally, no data sent to external servers

Probe Enterprise

Contact for pricing

Everything in Open Source, plus multi-repository architecture support, workflow automation, and team collaboration features.

  • Multi-repository architecture — Query across your entire system of services, not just one project
  • System-wide dependency awareness — Understand how services connect, which changes break what
  • Auto-generated onboarding docs — Personalized guides for every new hire, every repo, every team
  • Team knowledge capture — When engineers leave, their knowledge stays behind
  • Slack/Teams integration — Ask questions from where you already work
  • SSO & RBAC — Enterprise authentication and role-based access control
  • Workflow automation — Pre-built workflows for onboarding, knowledge transfer, documentation
  • Usage analytics — Track onboarding velocity, knowledge gaps, common questions
  • Priority support — Dedicated support team, custom integrations, training
  • On-premises deployment — Runs entirely in your infrastructure for maximum security

How to evaluate Probe

We recommend a two-phase approach: first, validate the core technology with open source on a single project, then pilot the enterprise features with your next new hire.

Phase 1

Technical Validation

~10 minutes

Pick any of these and have something running before your next meeting. No account required.

~2 min

Add Probe to AI Coding Tools

Get enterprise-grade code understanding in Claude Code, Cursor, or any MCP-compatible tool. Auto-detects auth, works with any LLM API. One command to install.

You get: A specialized AI agent for code search and analysis — finds the right context and reduces wrong answers with bounded, structured retrieval.
AI code editor setup →
~5 min

Deploy a Codebase-Aware Slack Bot

Create a Slack bot that answers questions about your codebase. Your team can ask questions in Slack and get intelligent answers grounded in actual code — no context switching.

You get: A Slack bot your team can query about any codebase. Run locally to test, then deploy anywhere. Measure how many interruptions it prevents.
Full setup guide →
~10 min

Generate Onboarding Docs

Run Probe against your main repository and generate an architecture overview, setup guide, and pattern documentation. See what new hires would get on Day 1.

You get: Auto-generated onboarding documentation for your codebase. Compare it to what you currently give new hires (if anything).
Onboarding workflow guide →
Phase 2

Onboarding Pilot

2-4 weeks

Once you've validated the core technology, run a pilot with your next new hire to test the full onboarding workflow and measure impact.

1
Pick a new hire cohort

Choose your next 1-3 new engineers joining the team. Ideally engineers joining a complex system where onboarding is typically slow.

2
Setup session with Probe team

We'll help you connect your repositories, generate initial architecture documentation, and configure the knowledge base for your team's patterns.

3
Give new hires Probe access

On Day 1, give them access to Probe instead of (or alongside) traditional onboarding materials. Let them ask questions, explore the codebase, and learn at their own pace.

4
Measure the difference

Track time-to-first-PR, number of questions asked to teammates, onboarding satisfaction scores. Compare to your historical baseline.

Success criteria: Measurable reduction in time-to-first-PR and questions-to-teammates. New hires report higher confidence and faster understanding. Senior engineers reclaim interrupted time.

Want to discuss how a pilot would work for your team?

Schedule a Technical Discussion

What engineering managers ask us

How is this different from internal wikis or Notion docs?

Wikis go stale the moment they're written. Engineers forget to update them when code changes. Six months later, they're lies.

Probe generates documentation directly from code. When the code changes, the documentation stays accurate because it's derived from the source of truth. Plus, it understands your codebase semantically — it can answer questions wikis can't, like "which services depend on this API?" or "how does data flow through the payment system?"

Will this actually reduce interruptions to senior engineers?

Yes, measurably. Teams report 50-70% reduction in onboarding-related Slack messages after deploying Probe. New hires get answers instantly instead of waiting for a teammate to respond. Senior engineers reclaim 5-10 hours per week previously spent answering "where is this?" questions.

The key is that Probe answers the questions accurately enough that new hires don't need to follow up. It's not just faster — it's self-sufficient.

What if our codebase is a mess with no documentation?

That's exactly where Probe shines. Legacy systems with zero documentation are where onboarding is hardest and where Probe has the biggest impact. It doesn't need documentation — it reads the code itself, understands structure, traces dependencies, and builds knowledge from what actually exists.

In fact, teams with poor documentation see the most dramatic improvements because the baseline is so low.

How long does it take to see results?

Immediate. The open-source version works out of the box — install it, point it at a repo, start asking questions. For enterprise deployments with multi-repo setup, initial configuration takes a few hours, then you're live.

The real measurement is your next new hire. Compare their ramp-up time to the previous cohort. Most teams see 40-60% reduction in time-to-first-PR within the first pilot.

Ready to transform your onboarding?

Let's talk about how Probe can help your new engineers ship code faster and reclaim your senior engineers' time. We'll show you how it works on your actual codebase and discuss how to structure a pilot.