Product / Features
See the work.
Understand the evidence.
Choose the next move.
A closer look at 48 capabilities across the Binomial engineering workspace: what you can inspect, the questions you can investigate, and the context that stays attached.
Start with GitHub in the hosted trial, or deploy the licensed source code in your own environment. Available evidence depends on your connected sources and setup.
Availability: The GitHub Cloud hosted trial is available. AI features and live Jira/Bitbucket connectors are not yet activated. Confirm advanced analysis and connection requirements for your deployment.
01 / Connect and scope
Choose the evidence that enters your workspace.
Connect the sources you use, select the repositories that matter, and follow collection as it runs.
Provider permissions and connection methods affect which evidence can be collected.
Compare integrations and setup optionsRead-only GitHub connection
Connect through the GitHub App and discover the repositories your installation can access. The hosted trial starts with GitHub Cloud.
Additional Git providers
GitLab and Git-compatible hosts have credential-based setup paths. Bitbucket setup is present, but its live connector is not yet activated. Confirm network access, repository scope, and collection depth for your deployment.
Repository selection and exclusions
Choose which repositories enter analysis and keep others outside the selected scope. Review newly discovered repositories as your installation changes.
Repository uploads and offline evidence
Bring a repository snapshot or supported offline export into a scoped analysis. API activity and continuous workflow history depend on the evidence supplied and any linked connector.
Historical collection
Collect available repository history through background jobs. Inspect progress and coverage while backfills build the evidence behind your reports.
Refresh and analysis controls
Request collection or analysis for supported sources. Full repository refresh currently applies to GitHub connectors and direct uploads; work runs in the background before reports become available.
02 / Delivery and activity
Follow the work from change to delivery.
Look beyond a single activity count. Connect commits, pull requests, reviews, validation, and delivery context.
Activity is a starting point for investigation; volume alone does not establish quality, productivity, or AI impact.
Activity over time
Explore 7-, 30-, 60-, 90-, and 180-day windows. Compare commit, pull-request, review, workflow, deployment, and lines-of-code activity where evidence is available.
Pull-request flow
See opened, reviewed, merged, and closed-unmerged pull requests. Inspect cycle time and merge-time signals alongside the work included in the window.
Review turnaround
Examine first-review timing, review activity, and the path to merge. Use the surrounding change context to investigate where work is waiting.
Checks and workflow context
Review collected workflow and check evidence around a change. Link validation runs to a pull request by its number or commit, where those relationships are available.
Connected change episodes
Inspect a change with its commits, reviews, workflows, checks, and repository context together. Episodes show when the supporting evidence is limited.
Release and deployment evidence
Include persisted release and deployment events in the delivery story. Coverage depends on the source events your connected systems expose.
03 / Codebase intelligence
Investigate the code behind the pattern.
Move from a repository-level question into indexed code, documentation, symbols, and relationships.
Index depth and relationship precision depend on the language, parser, collected source, and analysis state.
Code and documentation search
Search indexed source files and repository documentation. Keep the repository and source reference attached as you investigate a result.
Symbol and entity lookup
Find indexed symbols and entities, then retrieve their surrounding context. Use those references to understand where a concept appears in the repository.
Dependency-impact views
Explore upstream and downstream relationships with relevant risk context. Inspect the nodes and edges that support an impact question.
Repository relationship graph
Navigate repository-local structural relationships in an indexed graph. Select the repositories included in graph analysis and follow processing status.
Versions and pattern comparisons
Inspect available repository versions and compare indexed variants or patterns. Review the source references behind similarities and differences.
Evidence packs
Gather relevant repository evidence into a pack for a focused investigation. Retrieve the pack with its context rather than starting the same evidence search again.
04 / Evidence-backed workflows
Turn a question into a reviewable investigation.
Use task-oriented engineering analysis over persisted repository evidence, with structured context, evidence sections, freshness, and citations available for review.
AI features are not yet activated. Confirm availability of these evidence workflow modes for your deployment; missing or stale sources can limit the analysis.
Explain
Build context around code or a repository topic. Review a structured summary and focus questions anchored in the available evidence.
Assess
Frame a technical assessment around selected evidence and risk signals. Inspect the supporting context before deciding what needs deeper review.
Debug
Bring relevant files, symbols, and dependencies into a debugging investigation. Use the focus questions to guide the next engineering check.
Remediate
Prepare remediation context and suggested next actions from the evidence. Your engineering team reviews and carries out the change.
Govern
Organize evidence for a governance question with relevant repository context. Review the findings and actions within the limits of the collected sources.
Saved workflow reports
Return to saved analyses with their evidence sections, citations, next actions, and creation time. Check report freshness before reusing a conclusion.
05 / Portfolio and coaching
Move from the portfolio into the working context.
Help leaders and authorized reviewers understand patterns across repositories and individual contribution evidence.
Named-person reports require an explicit administrator grant. These signals support coaching and contextual review.
Review access and evidence boundariesEngineering portfolio overview
Review repository coverage and engineering activity together. Follow a portfolio question into the repositories and available evidence behind it.
Repository-level reports
Inspect repository activity, risk drivers, and supporting events in a focused report. Compare the interpretation with the scope and collection state.
Individual contribution dimensions
With explicit report access, inspect delivery, review, activity, ownership, and code-forensics dimensions. Keep the context of the work alongside the metrics.
Repository and event drill-down
Follow a contribution signal into the repositories and activity events that support it. Review the source evidence behind the aggregate view.
Coverage and confidence
See metric coverage and confidence tiers from high to insufficient. Confidence notes explain the available evidence and why a view may be incomplete.
Methodology in context
Inspect calculation and methodology metadata beside the report. Use the evidence to prepare a coaching conversation or a focused team experiment.
06 / Executive decisions
Connect engineering friction to investment choices.
Bring risk concentration, delivery friction, and modeled remediation exposure into the same leadership conversation.
Economics is supporting analysis. Repository activity by itself does not prove AI usage, spend, or causal return.
Explore the evidence walkthroughExecutive overview
Review top risks, delivery and review friction, modeled exposure, and recommended interventions. Use the overview to choose which repository or issue to investigate next.
Repository hotspots
Locate repositories where risk signals concentrate. Inspect the risk label, primary drivers, and supporting evidence behind a hotspot.
Modeled remediation exposure
Explore a monetary estimate derived from engineering evidence and model assumptions. Treat it as modeled exposure rather than observed financial performance.
Delivery and review friction
Bring workflow, validation, and review signals into the investment discussion. Examine what is driving the pattern before choosing an intervention.
Recommendations with context
Review suggested interventions alongside their confidence, limitations, and provenance. Bring team knowledge into the decision about what to fund or change.
Model and AI-signal transparency
Inspect model or formula versions and the state of AI evidence. Direct AI usage and spend claims depend on appropriate provider evidence being available.
07 / Confidence and provenance
Know what supports an interpretation.
Keep source state, completeness, freshness, and limitations visible as you move from evidence to a finding.
Confidence changes with permissions, collection completeness, source freshness, and analysis coverage.
Read the methodologySource references
Follow surfaced source links and evidence references back to the relevant event or repository context. Observed timestamps help establish when the evidence was captured.
Collection coverage
Inspect selected scope, object counts, and collection watermarks. See what has been collected and where the available history remains incomplete.
Pipeline readiness
Distinguish source collection, code synchronization, code analysis, and report readiness. Identify the stage that needs attention before expecting a usable report.
Freshness states
See whether a view is ready, stale, pending, limited, or unavailable. Refresh required evidence before reusing an analysis that is no longer current.
Evidence limitations
Review missing inputs and partial or blocked collection states. A limited finding stays distinguishable from one backed by broader evidence.
Observed and modeled context
Keep collected facts separate from their modeled interpretation. Provenance and confidence labels help reviewers judge what a conclusion can support.
08 / Access and operations
Give the right people a usable workspace.
Control customer access and give authorized administrators and operators the context to diagnose collection and analysis.
Available controls depend on the user role, source connection, and deployment configuration.
Choose your deploymentTenant-scoped workspaces
Customer requests resolve within the signed-in tenant context. Repository evidence and reports are scoped to that workspace.
Role-controlled setup
Owner and administrator roles govern sensitive setup and connection actions. Workspace membership controls who can enter the customer environment.
Explicit individual-report grants
Assign or revoke named-person report access separately. Members without that grant receive permitted aggregate views with restricted contributor detail withheld.
Connector and mirror diagnostics
Authorized operators can inspect connector state, mirror health, raw-data objects, and pipeline diagnostics. Use these signals to locate collection or processing problems.
Background processing
Collection, repository synchronization, and analysis run through background work. Dashboards read prepared workspace evidence as processing completes.
Retry and recovery visibility
Inspect failed or stale stages and use supported refresh or rebuild actions. Follow status after a retry to confirm the pipeline reaches the required state.
Run it in your environment
Source code included.
Your infrastructure.
Purchase the licensed source code and Dockerized distribution, then configure the integrations for your environment. Your team manages hosting, operations, and updates.
Optional expert assessment and coaching are available separately.
Discuss your requirementsStart with your engineering evidence
A clearer picture.
A better next move.
Connect your engineering sources and start understanding the patterns behind your outcomes. Try the hosted version with GitHub, or get the source code and deploy Binomial in your own environment.