FixDoc — Three intelligence engines for infrastructure fixes
Three Intelligence Engines // Infrastructure
Your infrastructure has a memory problem.
FixDoc gives it three brains.
FixDoc classifies every error by memory-worthiness, analyzes change impact with<br>score explanations, and learns from apply outcomes — so your team builds on<br>past fixes instead of repeating them.
Run the Demo
View on GitHub
Capture. Classify. Predict.
pipx install fixdoc && fixdoc demo tour
Local-first by default. Stores fixes in ~/.fixdoc. Git sync is optional. Private fixes never sync.
fixdoc — defer-first capture
// INSTALL<br>Get started in 60 seconds
Install FixDoc and start building your fix library.
quick start
Copy
Run the Demo
View on GitHub
// INTELLIGENCE ENGINES<br>Three intelligence engines
Each engine handles a different phase of your infrastructure lifecycle
Change Intelligence
Understands what a Terraform change will do before you apply it.
Change impact scoring with sigmoid formula (0–100)
BFS propagation through dependency DAG
IAM sensitivity scoring with wildcard trust detection
Smart fix matching with 8 change domains
Contextual checks from attribute + history
Score explanation bullets (action, IAM, impact, history)
AI-powered natural language explanations
Outcome learning with plan fingerprint linking
Failure Intelligence
Classifies, diagnoses, and routes errors when your commands fail.
Memory-worthiness classifier (memory-worthy vs self-explanatory)
AI error diagnosis explaining WHY it happened
Fix suggestions with effectiveness scores
Slack push notifications with Block Kit
Recurrence-based promotion (3+ similar → memory-worthy)
Memory Engine
Stores, classifies, and retrieves fixes with type-aware rendering.
4 memory types: fix, check, playbook, insight
Type-aware preview rendering in suggestions
Effectiveness tracking (applied_count, success_count)
Source error ID linking across sessions
Bidirectional markdown round-trip for Git sync
// CHANGE IMPACT ANALYZER<br>See your change impact. Understand why.
Understand how one change propagates through your infrastructure
Score:<br>/ 100<br>[HIGH]
Plan Change<br>Policy / RBAC<br>Workloads<br>Resources
aws_iam_role.app
DELETE
iam_policy_attachment<br>depth: 1
role_policy<br>depth: 1
lambda_function.api<br>depth: 2
ecs_task.worker<br>depth: 2
ec2_instance.web<br>depth: 2
s3_bucket.data<br>depth: 3
dynamodb.sessions<br>depth: 3
sqs_queue.events<br>depth: 3
cloudwatch.alarms<br>depth: 3
+25 delete action<br>+18 IAM policy changed<br>+15 BFS depth 3<br>+14 2 prior fix matches
Replay Animation
analyze change impact before applying
# Analyze your change impact before applying
$ terraform show -json plan.tfplan > plan.json
$ fixdoc analyze plan.json
Score: 72.4 / 100 [HIGH]
Why this scored 72:
+25 delete action on IAM role
+18 IAM policy field changed
+15 BFS propagation depth 3
+14 2 prior fix matches
Relevant Past Fixes: aws_iam_role [high: error code: AccessDenied]
Contextual Checks: [attr] Verify IAM trust policy principals
// USE CASES<br>For every stage of your journey
Individual Engineers
Build your personal fix library. Stop googling the same error twice. Your solutions, always at your fingertips.
Teams
Shared knowledge base via Git sync. When one engineer solves it, everyone has access. No more duplicate debugging.
Onboarding
New engineers inherit your team’s fix history on day one. Reduce ramp-up time, accelerate productivity.
CI/CD Pipelines
Run change impact analysis on every PR. Gate merges by severity. Track outcomes post-apply.
Start building your fix library
// POSITIONING<br>Not another tool
FixDoc does one thing well
FixDoc is NOT
✕ An incident management SaaS with dashboards and alerts
✕ A service catalog portal you have to log into
✕ A generic documentation tool
✕ An AI wrapper that requires an API key to function
FixDoc IS
✓ Terminal-first capture that fits your existing workflow
✓ A searchable fix history that lives on your machine
✓ An intelligent system that classifies errors, analyzes change impact, and learns from outcomes
Get started in 60 seconds →
// PRIVACY & STORAGE<br>Your data, your machine
What’s stored?
Issue descriptions, resolutions, tags, and timestamps. Each fix is a simple JSON entry + a markdown file.
Where does it live?
Everything is stored locally in ~/.fixdoc/. No cloud service, no remote database.
Does it phone home?
No. FixDoc makes zero network requests by default. It’s a pure CLI tool with no telemetry or analytics.
What about AI features?
AI diagnosis and AI explain are fully optional. Your API key stays on your machine. No data is sent unless you explicitly use --diagnose or --ai-explain. FixDoc works perfectly without them.