ProductHow it worksPricingAbout

Engineering intelligence for the AI era

See how AI is actually changing your engineering team

Your team ships more code than ever. But can you tell where AI is creating real leverage, where it is masking shallow work, and who is actually making the engineering decisions that matter?

See how it works

No surveillance. Team-level insights. Evidence you can verify.

fleurapp.io/app/overview
Overview
How AI is changing your engineering system
AI Involvement
0%
up from 34%
Decision Evidence
0%
down from 58%
Post-merge Change
0%
up from 9%
PR Throughput
0
+23% MoM
Key Insights
High Impact
AI code without real review is causing 3x more rework
19% post-merge changes vs 6% when reviewers push back
23 PRs match this pattern→
Trend to watch
Review depth dropped 35% as AI usage doubled
AI 34% to 61%, code-changing reviews 48% to 31%

Output no longer equals contribution

AI changed how software gets made. But your metrics have not caught up.

Six months ago, you could read the important PRs, sit in on standups, and have a feel for how the team was doing. Today, AI tools have tripled PR volume. The signal is drowning in output. You know your team is busy. You do not know if they are getting better.

Before AI
Engineer writes code
PR, commit, ticket
Output is visible
Manager observes output
Performance judgment

Code output was a rough proxy for contribution. Imperfect, but it worked.

After AI
Engineer + AI
Massive code, PRs, tickets
Where is the judgment?the missing layer
Engineering value

Output is no longer a proxy for contribution. AI generates code, but who made the decisions?

The question is not how much code your team ships. It is who made the decisions, and whether those decisions are any good.

Inside Fleur

Six capabilities that turn your GitHub data into engineering intelligence.

Insights
AllHigh ImpactTrend
High Impact
AI code without review causing 3x rework
19% post-merge changes vs 6% when reviewers push back substantively
23 PRs match this pattern→
Trend
Review depth dropped 35% in core services
Code-changing comments down from 48% to 31% of all reviews
2 repositories→
Positive
Backend teams show strongest AI efficiency gains
Cycle time reduced by 31% with no measurable quality decline
cycle time -31%→
Insights that find patterns for you
Fleur surfaces 3 to 5 actionable patterns per week, not 50 metrics.
PR #1842
Observed
+482 lines · 2 review comments · 3 follow-up commits
Inferred
High AI · Low judgment · High rework risk
Confidence
Medium, 3 consistent signals
Every conclusion has evidence
Click any insight to see the PRs behind it.
AI Involvement
T1 Explicit12%
T2 Strong38%
T3 Inferred22%
T4 Unknown28%
AI involvement, not AI guesswork
Evidence tiers: explicit, strong, inferred, unknown.
Decision Evidence
42%
Substantive mod
61%
Deep review rate
18
Arch decisions
27%
Direct accept rate
Review depth stable despite 31% AI growth
Human judgment, not human scores
We show where decision evidence is strong or missing.
Engineering Outcomes
1.8d
3.1d
Cycle time
17%
9%
Rework
1.2
1.7
Review rounds
AI-involvedNo observed AI
Cycle time↓ 1.8d vs 3.1d
Rework rate↑ 17% vs 9%
Follow-up PRs↑ 11% vs 5%
Faster cycles, slightly more rework. Association, not causation.
Outcomes that matter
Cycle time, rework, follow-ups, segmented by AI. Association, never causation.
AI ROI
AI tools spend
CursorCopilotClaude Code
$48k/yr
for a 40-person team
Engineering leverage
Cycle time−31%
Rework ratestable
Review depthstable
Equivalent to ~3.2 extra engineers
ROI ≈ 6.4x
AI ROI, not just adoption
Tie your AI tool budget directly to engineering outcomes.

Simple, transparent pricing

No per-seat surprises. No surveillance fees. Cancel anytime.

MonthlyAnnualSave 20%
Starter
$19/ developer / mo

For small teams getting visibility back

Up to 10 developers

  • GitHub integration, unlimited repos
  • Weekly insights, up to 5 per week
  • AI involvement analysis
  • Decision evidence signals
  • Engineering outcomes, rework and cycle time
  • Evidence explorer, insight to PR
  • Email support
Most Popular
Team
$29/ developer / mo

For growing teams that need the full picture

10 to 200 developers

  • Everything in Starter, plus
  • Unlimited weekly insights
  • Team and repository analytics
  • PR detail with full inspectability
  • Weekly report, email and Slack
  • Custom time ranges
  • Export, PDF and CSV
  • Priority support
  • Developer self-insights coming soon
Enterprise
Custom

For organizations with advanced needs

200+ developers

  • Everything in Team, plus
  • SSO and SAML authentication
  • SOC 2 Type II, in progress
  • DPA and custom data agreements
  • Custom integrations, GitLab, Bitbucket, Jira
  • Dedicated success manager
  • 99.9% uptime SLA
  • Custom data retention
  • Developer self-insights roadmap

Compare plans

FeatureStarterTeamEnterprise
GitHub reposUnlimitedUnlimitedUnlimited
Weekly insights5UnlimitedUnlimited
AI involvement analysis✓✓✓
Decision evidence signals✓✓✓
Engineering outcomes✓✓✓
Evidence explorerBasicFullFull
Team and repo analytics—✓✓
PR detail inspectability—✓✓
Weekly report—Email + SlackCustom
Export—PDF + CSVCustom
API access——✓
Developer self-insights—Coming soonRoadmap
SSO and SAML——✓
SOC 2——In progress
SupportEmailPriorityDedicated

All plans include a 14-day free trial. No credit card required. Prices in USD. Annual billing saves 20%.

From GitHub to insight in four steps

Connect your repos. Fleur does the rest.

01
Connect GitHub

One-click OAuth. Select your repos. Fleur starts analyzing your history.

Connected
42 repos
02
Fleur analyzes

PRs, commits, reviews, comments. Every signal extracted and cross-referenced.

Analyzing...
847 PRs processed
03
Patterns emerge

AI involvement times human judgment times outcomes. Fleur finds the non-obvious.

High impactTrend
2 patterns found
04
You investigate

Drill from insight to repo to PR to diff. Every conclusion comes with evidence.

Insight→Repo→PR #1842
ObservedInferredConfidence
Why Fleur

Built for leaders who have lost their feel for the team

The old metrics stopped working the moment an AI agent could generate 500 lines before standup. Fleur gives you a new way to see.

I used to know every important PR in my org. Now we merge 200 a week and I cannot tell which ones were actually thought through. Fleur is the first thing that has given me that visibility back.

VP of Engineering80-person startup
Team-level first

Insights roll up to teams and repositories. Developer profiles coming next.

No surveillance

We do not track keystrokes, screens, or editor activity. Only your GitHub history.

Evidence you can verify

Every conclusion traces back to specific pull requests you can open yourself.

For teams of 20 to 500 engineers. SOC 2 Type II in progress.

Know what is actually happening in your codebase

Connect GitHub. Get your first insights in 15 minutes. No credit card.

Why I Built Fleur

Six months ago, my team started shipping three times more code. And I couldn’t tell if we were building more value, or just more volume.

AI has made producing software dramatically cheaper. PR volume is up, cycle times are changing, and teams are shipping more than ever. But there’s a paradox: we’re producing more, while engineers feel like they have less time to actually think.

The default response is to push harder on the same old metrics: more PRs, more tickets, more commits. Those metrics aren’t useless. They’re just measuring the wrong layer.

When software becomes cheap to produce, the scarce resource is no longer code. It’s human judgment.

I looked for tools that could help answer where that judgment was actually happening. Most either measured raw output or monitored individual AI usage. Neither told me how AI was changing engineering decisions, or whether those decisions led to better outcomes.

So I built Fleur.

Fleur starts with the evidence already in GitHub and connects the dots: AI-assisted work -> human judgment -> engineering decisions -> outcomes. We started at the team level because that’s where individual decisions compound into real results.

The question I’m trying to answer is simple:

In a world where output is no longer scarce, how do we know we’re actually building something valuable?

Fleur was founded in 2026. Based in New York.

Privacy Policy

Last updated: August 2026

What we collect

We collect data from your GitHub account that you authorize us to access. This includes pull request metadata, commit information, review comments, and repository structure. We do not collect source code content beyond what is visible in pull request diffs. We do not track IDE activity, keystrokes, screen time, or personal browsing data.

How we use it

We use GitHub data to generate engineering insights and analytics for your organization. We do not sell your data. We do not use your data to train AI models. We start at the team and repository level, where engineering decisions actually happen.

Data retention

We retain your GitHub data for as long as your account is active. You can request deletion of your data at any time by contacting hello@fleurapp.io. Upon account deletion, all associated data is removed within 30 days.

Your rights

You can access, export, or delete your data at any time. You can disconnect GitHub integration from settings at any time. You can request a copy of all data we hold about your organization.

Contact

Questions about privacy? Email hello@fleurapp.io.

Terms of Service

Last updated: August 2026

Acceptance

By using Fleur, you agree to these terms. If you do not agree, do not use the service.

Service description

Fleur provides engineering intelligence analytics based on GitHub data. The service is provided “as is” without warranties of any kind.

User responsibilities

You are responsible for maintaining the security of your account. You must not use Fleur for employee monitoring, performance evaluation, or individual ranking. You must not attempt to reverse engineer the service or access data you are not authorized to view.

Data ownership

You retain ownership of your data. We claim no ownership rights to your GitHub data or derived analytics.

Pricing and billing

Paid plans are billed monthly or annually. You can cancel at any time. Cancellations take effect at the end of the current billing period. No refunds for partial periods.

Limitation of liability

Fleur is not liable for any indirect, incidental, or consequential damages arising from use of the service. Our total liability is limited to the amount paid in the preceding 12 months.

Changes to terms

We may update these terms from time to time. Material changes will be notified via email at least 30 days in advance.

Contact

hello@fleurapp.io

Security

Data encryption

All data is encrypted in transit using TLS 1.3 and at rest using AES-256. GitHub access tokens are encrypted at rest and never logged.

Access control

We use the principle of least privilege. GitHub OAuth access is limited to read-only permissions for pull requests, commits, and repository metadata. We cannot modify your repositories or source code.

Infrastructure

Our infrastructure runs on major cloud providers with SOC 2 Type II certified data centers. We are currently working toward our own SOC 2 Type II certification.

No employee monitoring

Fleur starts at the team and repository level, where engineering decisions actually happen.

Responsible disclosure

If you discover a security vulnerability, please report it to security@fleurapp.io. We respond to all reports within 48 hours.

Get early access

We are onboarding design partners now. Leave your email and we will reach out to schedule a 15-minute walkthrough.