Overview
Session outcomes, value, and proof from your recent Avorelo work.
History
Outcome receipts from recent sessions. What was found, fixed, saved, and proved.
📋
No session history yet
Session receipts appear here after you run Avorelo in a project. Each receipt shows what was found, fixed, saved, and proved.
npx -y avorelo@latest activate --scope project-wide --claim <your_token>
Settings
Workspace, privacy, permissions, and behavior configuration.
Workspace
Active workspace
Not activated yet
AI coding tools detected
Run activation to detect tools
Defaults
File access scope
Project-wide by default. All project files are accessible to Avorelo during sessions.
Proof stays local
Session proof, logs, and receipts remain on this machine.
⚙
More settings coming soon
Settings for background operation, context refresh, analytics, permissions, safe fixes, and visual QA will be configurable in a future release. Currently, Avorelo uses safe defaults that work for most projects.
Background operation
Analytics
Permissions
Safe fixes
Visual QA
Cloud sync
Plan
Current plan
Free
Account
Your plan, billing, and account details.
Account
Email
Not connected
Current plan
Free
Billing period
N/A (Free plan)
Cloud sync
Off (proof stays local)
Usage this month
Sessions
—
No data yet
Tokens avoided
—
No data yet
Time saved
—
No data yet
Support
Diagnostics, bundles, and help resources.
Create support bundle
Package evidence summaries and configuration state for a support request. Raw prompts, raw code, and secrets are excluded by default.
View local diagnostics
Check Avorelo integration status, detected tools, workspace state, and configuration health locally on this machine.
Integration status
Check whether Avorelo can see your AI coding tools, workspace, and MCP configuration correctly.
Documentation and contact
Read the docs or get in touch with the Avorelo team directly.
Contact support
Support bundle privacy: A redacted bundle includes evidence summaries and configuration state. It does not include raw prompts, raw code, or secrets by default.
Team Reports
Evidence-backed outcome reports for AI-agent work across your team. Teams plan
This report is available on the Teams plan.
Get Teams access
Teams plan
AI Spend & Proof
Summary
Shows whether AI usage produced evidence-backed work, not just token activity.
Confidence: Estimated for cost metrics
184k
Tokens avoided
12
Validated AI workflows
3
High-cost without proof
$14.20
Est. cost avoided
Confidence: Estimated
What this means
This shows whether AI usage produced evidence-backed work, not just token activity.
Why it matters
Teams do not need more AI usage. They need to know which AI work created value and which workflows are wasting budget or context.
How Avorelo calculates it
Avorelo compares repeated context, model route, task type, proof state, and estimated token/cost impact.
What to review
High-cost workflows without validation, repeated setup, and expensive model usage for low-risk tasks.
Evidence
Evidence links attach to proof receipts, cost estimates, repeated context reduction, and validation signals.
Trend
Track whether validated AI work is increasing while repeated context and unsupported high-cost work are decreasing.
Recommended action / policy
Move low-risk tasks to cheaper routes, require proof for expensive workflows, and reduce repeated context before it reaches the model.
Customer-facing article
From AI Spend to AI Proof: How to Know What Your AI Budget Actually Produced →
Privacy note
Reports are aggregated by repo, workflow, tool, model class, capability, risk, and proof type. They do not expose raw prompts, raw code, or individual ranking.
This report is available on the Teams plan.
Get Teams access
Teams plan
What Got Done
Summary
Separates AI activity from validated outcomes.
Confidence: High for validated outcomes
31
AI-assisted work units
14
Validated outcomes
5
Reusable artifacts created
6
Workflows ended without validation
What this means
This separates AI activity from work that actually produced a validated outcome.
Why it matters
Prompt counts and agent runs do not prove progress. Teams need to know what was completed, validated, reused, or left unfinished.
How Avorelo calculates it
Avorelo groups AI sessions into work units and connects them to proof signals such as tests, receipts, visual evidence, safe-fix receipts, or handoff evidence.
What to review
Workflows with high activity and low proof, repeated unfinished runs, or outputs that were not validated.
Evidence
Evidence links attach to receipts, validation output, visual evidence when relevant, and handoff evidence.
Trend
Track whether validated outcomes and reusable artifacts are increasing while unfinished AI activity decreases.
Recommended action / policy
Require proof for critical workflows, improve handoff rules, and standardize successful work patterns.
Customer-facing article
AI Activity Is Not Productivity: Measuring What Actually Got Done →
Privacy note
Reports are aggregated by repo, workflow, tool, model class, capability, risk, and proof type. They do not expose raw prompts, raw code, or individual ranking.
This report is available on the Teams plan.
Get Teams access
Teams plan
Scope & Safety
Summary
Shows where AI agents tried to go broader than intended, and where Avorelo narrowed or blocked the action before involving a human.
Confidence: High for scope repair records
8
Scope drifts repaired
126 to 19
Files narrowed
4
Risky actions blocked
11 avoided / 3 needed
Approval interruptions
What this means
This shows where AI agents tried to go broader than intended, and where Avorelo narrowed or blocked the action before involving a human.
Why it matters
The problem is not just risky AI. The problem is review fatigue. If every drift becomes an approval, users start ignoring approvals. Avorelo repairs scope first and asks only when real authority is needed.
How Avorelo calculates it
Avorelo compares the intended task scope with planned or attempted actions, then records scope repair, blocked actions, reduced permissions, and remaining approval decisions.
What to review
Sensitive changes that still require authority after scope repair, repeated scope drift in the same workflow, or capabilities that request too much access.
Evidence
Evidence links attach to scope repair receipts, blocked action records, narrowed file/action sets, and approval decisions.
Reduced one agent run from 31 files to the 2 requested files before approval.
Trend
Track whether scope drift, risky actions, and remaining approvals are decreasing over time.
Recommended action / policy
Tighten defaults for workflows with repeated drift, reduce broad access patterns, and require stricter approval only for sensitive actions that remain after repair.
Customer-facing article
Scope Repair Before Approval: How to Stop Babysitting AI Agents →
Privacy note
Reports are aggregated by repo, workflow, tool, model class, capability, risk, and proof type. They do not expose raw prompts, raw code, or individual ranking.
This report is available on the Teams plan.
Get Teams access
Teams plan
Workflow Health
Summary
Shows whether the team's AI workflows are getting cleaner or noisier over time.
Confidence: High for detected signals
9
Stale instructions found
6
Repeated setup blocks reduced
3
Noisy capabilities controlled
4 stopped / 2 need review
Failed loops
What this means
This shows whether the team's AI workflows are getting cleaner or noisier over time.
Why it matters
AI work gets slower and riskier when instructions accumulate, tools overlap, context becomes stale, and agents repeat the same mistakes.
How Avorelo calculates it
Avorelo looks for repeated setup, stale or conflicting instructions, noisy capabilities, failed loops, duplicated workflow steps, and capabilities that lack enough trust/evidence.
What to review
Capabilities that should be disabled, repeated loops, stale project knowledge, and noisy instructions that keep increasing context load.
Evidence
Evidence links attach to repeated setup receipts, stale instruction findings, failed loop records, capability hygiene results, and tool trust signals.
Trend
Track whether repeated setup, stale instructions, failed loops, and noisy capabilities are decreasing.
Recommended action / policy
Retire stale instructions, quarantine noisy capabilities, standardize useful practices, and reduce duplicated workflow steps.
Customer-facing article
Why AI Workflows Get Messy and How to Keep Them Clean →
Privacy note
Reports are aggregated by repo, workflow, tool, model class, capability, risk, and proof type. They do not expose raw prompts, raw code, or individual ranking.
This report is available on the Teams plan.
Get Teams access
Teams plan
Context & Continuity
Summary
Shows whether the team is repeatedly re-explaining the same work to AI, or preserving useful context that helps future sessions start faster.
Tokens: Estimated
18
Repeated explanations avoided
7
Reusable context packets preserved
5
Stale context blocks filtered
11 captured / 42k tok saved
Handoffs
Tokens: Estimated
What this means
This shows whether the team is repeatedly re-explaining the same work to AI, or building reusable context that helps future sessions start faster.
Why it matters
A lot of AI waste comes from rediscovery. Teams lose time and tokens when every session starts from zero or carries stale context forward.
How Avorelo calculates it
Avorelo detects repeated setup, reused task context, stale context filtered from future runs, and handoff/receipt evidence that can safely support continuation.
What to review
Workflows with repeated setup, stale handoffs, conflicting project knowledge, or context that is too broad for the task.
Evidence
Evidence links attach to handoff receipts, reusable context records, stale context filtering, and re-entry evidence.
Trend
Track whether repeated explanations go down and reusable context, handoffs, and re-entry readiness improve.
Recommended action / policy
Improve handoff defaults, reduce broad context carry-forward, standardize reusable context, and remove stale project knowledge.
Customer-facing article
Stop Re-Explaining Your Project to AI →
Privacy note
Reports are aggregated by repo, workflow, tool, model class, capability, risk, and proof type. They do not expose raw prompts, raw code, or individual ranking.
Needs team review
Only real decisions appear here after Avorelo has already cleaned, narrowed, or blocked what it safely can. Teams plan
Agent requested write access to 14 files outside declared scope
2026-05-31 · 14:19 · avorelo / apps / public-web
What happened
During a refactor session the agent planned changes across 14 files, including 9 outside the declared scope boundary for the task.
What Avorelo already fixed
Narrowed the write boundary from 14 to 5 files. 9 files were removed from the plan before execution. Scope drift repaired before escalation.
Why review is still needed
The remaining 5 files include a shared config file used by other repos. Writing to it without team review could affect other projects.
Evidence
Scope repair receipt attached · 9 file narrowing logged · Config file cross-repo usage detected
Agent requested shell command outside normal capability scope
2026-05-30 · 16:44 · avorelo / infra
What happened
An agent running a deploy task requested access to run a destructive database migration command not included in its declared capabilities.
What Avorelo already fixed
Blocked the command at the capability boundary. The deploy continued with the allowed steps. Only this migration decision was escalated.
Why review is still needed
The migration is a real task that needs to run, but requires explicit authorization because it is destructive and outside the declared capability for this session.
Evidence
Blocked action record attached · Capability boundary logged · Deploy receipt shows partial completion pending this decision
Team Settings
Team-level defaults managed by the team admin. Applies to all team members.
Team Settings require the Teams plan. Contact us to get access.
Get Teams access
Workspace
Default scope
Default write boundary applied to all AI coding sessions for team members
Proof path
Where proof receipts are stored. Default: .avorelo/receipts/
Privacy
Local only
Session proof, logs, and receipts remain on each member's machine unless cloud sync is enabled
Telemetry
Share anonymous usage signals to improve Avorelo. Off by default for Teams.
Team defaults
Reporting period
Period used for team report aggregation
Data retention
How long proof receipts and session data are kept
Aggregation level
Reports are aggregated at this level. Never by individual employee productivity rank by default.
Allow member report viewing
Whether individual team members can view team reports
Permissions
Auto route
Automatically route AI tasks to the appropriate model and scope
Require approval above
Risk level at which AI actions require human approval
Approvals
Approval threshold
Team-wide threshold for when Avorelo escalates to human review
Evidence & exports
Safe fixes
Allow Avorelo to apply safe cleanup automatically for team members
Safe fix types
Which fix categories are applied automatically
Export receipts
Export team proof receipts for compliance or review
Plan
Current plan
Teams