SR&ED operating system for evidence-first teams
A practical workspace for preserving technical facts, source evidence, approvals, and advisor-ready drafts while your startup is still building.
Built for the founder who cannot afford a year-end archaeology project
Most early-stage software teams know the hard technical work when it happens. By filing season, the source links, eval failures, tradeoffs, and approvals are scattered across tools. This repo gives the team a small operating rhythm before that context disappears.
During the sprint
Capture technical uncertainty, experiments, failures, model evals, decisions, and the evidence that proves when they happened.
During the month
Review what is accepted, excluded, weak, blocked, sensitive, or waiting on finance, privacy, CTO, or advisor input.
Before advisor review
Hand over a clean binder with source-backed narratives, cost support, privacy decisions, approval status, and unresolved questions.
Start with the company, then let evidence drive the story
Confirm the company and claim period
Fill in the company profile, fiscal period, owners, policy refresh status, and human approval gates.
company-profile.mdMap systems of record before scanning
List the repos, tickets, docs, eval systems, payroll, accounting, and customer evidence restrictions that matter.
inventory-runbook.mdRun the founder and CTO intake
Turn hard technical work, failed paths, customer context, and routine exclusions into reviewable notes.
first-time-founder-intake.mdTrack projects, uncertainties, experiments, and AI evals
Connect each technical unknown to attempts, tests, datasets, metrics, failure analysis, and dated source artifacts.
experiment-ledger.mdReview people, finance, privacy, and exclusions
Keep costs, locations, grants, cloud usage, customer data, sensitive IP, and routine work visible for human review.
finance-evidence-ledger.mdBuild the binder from approved facts only
Package the claim period for a CPA or SR&ED specialist without hiding weak evidence or unresolved advisor questions.
final-review-binder.mdA source-linked workspace that can survive scrutiny
Each artifact exists to answer a question the claim process will eventually ask. The repo keeps facts, interpretation, approvals, and sensitive raw material in separate lanes.
Claim question index
The founder control board for what is answered, blocked, excluded, or waiting for human review.
Evidence index and cards
Dated artifacts from Git, tickets, evals, docs, meetings, logs, finance, and payroll stay traceable.
Experiment ledger
Hypotheses, attempts, tests, results, and conclusions stay tied to the technical unknown being investigated.
AI eval register
Models, prompts, datasets, baselines, thresholds, metrics, traces, and failure patterns are captured while fresh.
Time and finance ledgers
People, contractors, cloud, model API, grants, assets, locations, and source documents are mapped for review.
Final review binder
An advisor-facing index of approved facts, draft narratives, evidence, costs, privacy status, and open questions.
Useful evidence support, not tax advice
This operating system helps a startup preserve what happened and prepare materials for qualified review. It does not decide final eligibility, calculate a credit, promise a refund, file a claim, or replace a CPA or SR&ED specialist.
View 14-day Setup Dashboard