This is the full developer documentation for Aevy docs # Activity log and comments > Every record keeps a history of what happened, and a place to talk about it. Every record keeps a history of what happened, and a place to talk about it. ## What the activity log is [Section titled “What the activity log is”](#what-the-activity-log-is) Most records in Aevy carry an Activity section: a running history of what happened to them, newest at the top, each entry naming who did what and when. A comment box sits at the top of the same feed, so the discussion lives next to the history rather than in a separate thread. You can see how a task, a finding, or a [document](/documents/) reached its current state, and talk about it in place.  The activity feed, expanded: the comment box on top, then every change with who made it and when. Aevy’s own actions show like anyone else’s. ## Where you will find it [Section titled “Where you will find it”](#where-you-will-find-it) The same Activity section appears across the platform and works the same wherever you are. You will find it on: 1. **[Tasks and reminders](/reminders/), [findings (punchlist)](/punchlist/), and [events](/events/).** The work items. 2. **[Systems](/systems/), contacts, and companies.** 3. **Documents and requirements.** 4. **An asset’s overview**, and inside the quick-view drawer you open from any table row. If a record has a history worth keeping, it has an activity log. ## What gets logged [Section titled “What gets logged”](#what-gets-logged) The feed records the changes that matter, and nothing happens silently: 1. **Creates and deletes.** “created the task”, “deleted the system”. 2. **Field changes.** “set financial year start to …”, or “changed status from … to …”, with a status shown as its coloured badge. 3. **Assignments** and links to documents and other records. 4. **Aevy’s own actions.** Accepting a [suggestion](/suggestions/), extracting a fact, posting a comment. When Aevy acts, it is in the log like anyone else. ## Comments [Section titled “Comments”](#comments) Write a comment in the box at the top of the feed. Comments are rich text and take @ mentions, so you can point at a record or bring in a teammate (see the [mentions guide](/mentions/)). You can reply to a comment, react to one (a thumbs up, an eye, a hundred), and edit or delete your own. Mention a colleague to loop them in, or mention [Squirrel](/squirrel/) to ask it to act on the record. ## Getting notified [Section titled “Getting notified”](#getting-notified) Use Notify me on a record to subscribe to its updates, and choose which kinds of change reach you: a status change, an assignment, a new comment. The subscription is per record, so you follow the handful you care about without hearing about everything. # Compliance and obligations > Track everything your contracts require, and see where you are not compliant. Track everything your contracts require, and see where you are not compliant. Beta Compliance is a beta feature. Its structure and naming are still changing, so expect the screens and the words on them to shift over time. In this guide we call the tracked items obligations; some screens still label them “requirements” for now. ## What compliance tracks [Section titled “What compliance tracks”](#what-compliance-tracks) An obligation is something your contracts require you to do or to hold: a document to keep on file, a report to file on time, an inspection to run, a certificate to renew. The [Compliance page](https://aevy.app/dashboard/requirements) gathers these across your assets and tracks each one. When an obligation is not met by the time it is due, that is a non-compliance, and the page makes it visible so nothing slips quietly. ## Where obligations come from [Section titled “Where obligations come from”](#where-obligations-come-from) You do not build the list by hand. Aevy reads your contracts and extracts the obligations in them automatically, then tracks each one on the cadence the contract defines. A one-off deliverable is tracked once; a recurring duty (an annual report, a periodic inspection) repeats on its own schedule. Because the obligations come from the contracts you actually signed, the list reflects your real commitments rather than a checklist someone remembered to write. ## Reading the page [Section titled “Reading the page”](#reading-the-page) Obligations are grouped by lifecycle phase (pre-construction, construction, operations) and by category within each (grid, offtake, permitting, and so on). A health bar at the top shows how many are on track (for example, 40 percent on track, 77 of 192), and each group in the list carries its own overdue count, so you can see at a glance where the pressure is. The tabs separate Active obligations from Drafts and Archived, and you can search, [filter](/filtering/), and regroup the list.  The Compliance page: obligations grouped by phase and category, with an on-track health bar. ## Anatomy of an obligation [Section titled “Anatomy of an obligation”](#anatomy-of-an-obligation) Each obligation is a card. It names the asset it belongs to and what is required, and around that it carries the three things you check at a glance: 123 One obligation for Sunny Solar. The numbers map to the parts below. 1. **Due date and cadence.** When it is due, and how often it repeats. This one is one-time; a recurring obligation shows the schedule its contract sets. 2. **How it is checked.** The check Aevy runs to decide whether it is met, here a [document gap analysis](/documents/): is the required document present? 3. **Status.** Met or not. A count with a warning, like 0 of 1, means the obligation is not yet met, which is a non-compliance. ## Non-compliance [Section titled “Non-compliance”](#non-compliance) When an obligation is not met, it stands out. The card carries a warning, the item counts toward the overdue total on its group, and it pulls the on-track health bar down. So a non-compliance is never buried in a list: you see it on the obligation, on its category, and on the asset as a whole, which is exactly where you would look to judge your exposure. # Core concepts > The things you work with in Aevy: what they are, how they fit together, and where you meet them. The things you work with in Aevy: what they are, how they fit together, and where you meet them. How to read this Read it top to bottom the first time; after that, treat it as a reference. Almost everything in Aevy hangs off an asset, so that is where we start. For a one-line lookup, the [glossary](/glossary/) holds the same concepts as a table. ## Start with your asset [Section titled “Start with your asset”](#start-with-your-asset) ### Asset [Section titled “Asset”](#asset) An asset is a single renewable energy project: a solar plant, a wind farm, a storage site, an EV charging site, or a combination of those. It is the home base everything else attaches to: its documents, its equipment, its work, its people. The Assets section of the [sidebar](/getting-around/) lists every asset you can see. ### System [Section titled “System”](#system) A system is a major technical part of an asset: a turbine, the substation, an inverter station. It holds the specifications and location for that equipment, and it is what inspections and defects roll up to. The [systems and components guide](/systems/) walks the views. ### Component [Section titled “Component”](#component) A component is a part inside a system: a gearbox, an inverter, a rescue kit. Components are the level where inspections and defects actually land, and they add up to their system’s state. Components are mapped to systems when your asset is set up, so the hierarchy is already in place when you arrive. ## Documents and what is inside them [Section titled “Documents and what is inside them”](#documents-and-what-is-inside-them) ### Document [Section titled “Document”](#document) A document is any file kept against an asset: contracts, permits, reports, manuals, inspection logs. Documents arrive through a sync with your source, so there is no upload step and nothing to file by hand. When one lands, Aevy reads it, classifies it, and pulls out its facts. The [documents guide](/documents/) covers finding and reading them. ### Classification [Section titled “Classification”](#classification) Three fields classify every document so you can filter to it: 1. **Type.** The top category: contract, report, design, notification, invoice, and so on. 2. **Subtype.** A refinement within the type: a contract might be an amendment or an appendix; a report might be a statutory inspection or an audit. 3. **Scope.** The subject area, which depends on the type: a contract is tagged O\&M, EPC, or PPA; a report might be legal, technical, or financial. A signature state alongside these tells you whether the document is signed, partially signed, or not signed. ### Extracted fact [Section titled “Extracted fact”](#extracted-fact) A fact is a single structured value Aevy pulled out of a document: a warranty start date, a base fee, an insured sum. There are several hundred kinds, and they turn a PDF into data you can filter, query, and ask [Squirrel](/squirrel/) about. Every fact keeps a link to the exact line it came from; the [lineage guide](/lineage/) shows how to reveal it. ### Related documents [Section titled “Related documents”](#related-documents) Documents rarely live alone. Aevy links versions and relatives of the same document: a newer version marks the older one Superseded, a copy is marked Duplicate, and amendments and appendices point at the agreement they belong to, each with a link to the active version. ### Pinned documents [Section titled “Pinned documents”](#pinned-documents) Pinning keeps a document one click away on its asset: the current contract, the latest inspection. Pinned documents gather on the asset and under the Pinned filter on the Documents page. The [favorites and pinning guide](/favorites/) covers it. ## What Aevy proposes [Section titled “What Aevy proposes”](#what-aevy-proposes) ### Suggestion [Section titled “Suggestion”](#suggestion) A suggestion is Aevy proposing a specific change to your data after reading a document: a component to add, a reminder to create, a property to set. A suggestion is always a draft: nothing changes until you accept it, and each one names the document and page it came from. The [suggestions guide](/suggestions/) covers the review loop. ### Confidence [Section titled “Confidence”](#confidence) Every suggestion carries a percentage: Aevy’s own estimate of how well the document supports it. Treat it as a prompt for attention, not a verdict: the score tells you where to look hardest, not what to decide. ## The work you track [Section titled “The work you track”](#the-work-you-track) Aevy keeps four kinds of work record, and they meet you in the same shape: a list per page, the shared [filter bar](/filtering/), and an activity feed on each record. 1. **Reminder.** A dated task on an asset: a calibration to run, a certificate to renew. See the [reminders guide](/reminders/). 2. **Punchlist item.** A defect or issue found on an asset, tied to the system or component it affects, with a severity and a status from report to resolution. See the [punchlist guide](/punchlist/). 3. **Event.** Something happening on an asset on a date, with a start, an end, and a status: a maintenance window, an inspection visit, a renewal. See the [events guide](/events/). 4. **Obligation.** Something a contract requires you to do or hold, tracked on the cadence the contract defines. An obligation is not a task itself: it defines what must be done and how often, and an unmet one is a non-compliance. See the [compliance guide](/compliance/) (beta; some screens still say “requirements”). ## People and companies [Section titled “People and companies”](#people-and-companies) ### Contact [Section titled “Contact”](#contact) A contact is a person tied to an asset or a company: a project manager, an O\&M lead, a manufacturer’s representative, with the details you need to reach them. Contacts have their own [page](https://aevy.app/dashboard/contacts) in the sidebar menu. ### Company [Section titled “Company”](#company) A company is an organization involved with an asset: an owner, an operator, an O\&M provider, a manufacturer, an insurer. Companies group the contacts who work there. ## Working together [Section titled “Working together”](#working-together) ### Comment [Section titled “Comment”](#comment) A comment is how you talk about a record without leaving it. Every record with an activity feed takes comments at the top of that feed, so the discussion sits next to the history. See the [activity log guide](/activity/). ### Mention [Section titled “Mention”](#mention) Type @ anywhere you write to name a specific record in a sentence: an asset, a document, a task, a teammate, or Squirrel itself. The mention points at the record itself. See the [mentions guide](/mentions/). ### Squirrel [Section titled “Squirrel”](#squirrel) Squirrel is the AI asset manager in the chat panel. It answers from your own data using the same tools you have, shows the steps it took, and cites the documents behind its answers. When it proposes work, the proposal arrives as a suggestion for you to accept or dismiss. See the [Squirrel guide](/squirrel/). ### Activity [Section titled “Activity”](#activity) Activity is the audit trail. Every meaningful change to a record is logged with who did it and when, and when Aevy acts, it is in the log like anyone else. Use Notify me on a record to hear about the changes you care about. ## Modes [Section titled “Modes”](#modes) A mode is the lens you work through, and it shapes the sidebar menu to fit the job: Asset Management for the day-to-day, Due Diligence for reviewing a deal, and Owner for the lighter ownership-level view. Switch modes from the organization name at the top of the sidebar; the [getting around guide](/getting-around/) shows where. # Documents > Find, read, and track every document on your assets, without filing anything by hand. Find, read, and track every document on your assets, without filing anything by hand. ## What the Documents page is [Section titled “What the Documents page is”](#what-the-documents-page-is) Every document connected to your assets lands on the [Documents page](https://aevy.app/dashboard/documents): inspection reports, contracts, drawings, operational reports. Documents arrive through a sync with your source: there is no upload step and nothing to file. The sync brings the folders too, not only the files. The structure you see is the structure at your source. When a document arrives we read it, pull out its key facts, and track where it is in processing, so you can find the right version without opening folders one by one. The page has two parts: a folder tree on the left, synced from your source, and a table on the right that lists documents once you narrow to what you want.  The Documents page. The table stays empty until you pick a folder, search, or add a filter. ## Finding a document [Section titled “Finding a document”](#finding-a-document) Four tools narrow the list, and they combine: 1. **Folders.** The tree on the left is your source structure, synced across: you browse the same folders you have at the source. You do not create or manage them here. Tick a folder to show only its documents. 2. **Search.** Search documents matches on document names and briefs. 3. **Quick filters.** The chips Assets, Pinned, Contracts, Reports, Designs, and Other jump to a common cut in one click. 4. **Add Filter.** For anything more specific, build a filter on any column. See the [table filtering guide](/filtering/) for how filters work. Saved views sit along the top (All documents, and any views your team has saved). Each view remembers its own folders, filters, and columns, so a shared cut is one click away.  A narrowed list. Here the quick action for recent uploads added a Created At is after filter; the table shows what matches. ## Reading a document [Section titled “Reading a document”](#reading-a-document) Open a row to read the document in full. The original opens in the main pane, with a details panel beside it: the document type, the parties, the page count, and its signature state (signed, partially signed, or not signed). Tabs group the related work, any requirements the document touches, and other versions of the same document, so you can move from the file to what depends on it without leaving the page. ## Document status [Section titled “Document status”](#document-status) Each document carries a status while we process it, so you know whether it is ready to trust. Most settle on Success. Along the way you may see OCR In Progress or Extraction Pending (we are still reading it), and occasionally a state that needs attention, such as Encrypted, Corrupt, or Document Too Long. When a newer version of the same document arrives, the older one is marked Superseded and a duplicate is marked Duplicate, each with a link to the active version. You can still open the old one, but the status tells you it is not the one to work from. ## Keeping a document close [Section titled “Keeping a document close”](#keeping-a-document-close) Pin the documents you return to often. A pinned document stays one click away under the Pinned filter and on its asset. See the [favorites and pinning guide](/favorites/) for how pinning works. # Events > Schedule and track activities on your assets, with dates, type, and status. Schedule and track activities on your assets, with dates, type, and status. ## What an event is [Section titled “What an event is”](#what-an-event-is) An [event](https://aevy.app/dashboard/events) is something happening on an asset on a date: a scheduled maintenance, a statutory inspection, a contract renewal, a recorded incident. Each event has a type, a start and an end, and a status, so the calendar of work across your portfolio is one list you can plan from.  The Events page. Each row is a dated activity on an asset, with its type, status, and dates. ## Finding an event [Section titled “Finding an event”](#finding-an-event) The list filters like every table in Aevy (see the [table filtering guide](/filtering/)). What is specific here: 1. **Saved views.** All events and My events are prebuilt cuts, each with its own count. 2. **Search.** Search events matches on the event title. 3. **Filter.** Filter and Add Filter narrow by any column, for example status or a date range. Each row names the asset, what is happening, the kind of event, its status from Scheduled to Confirmed, its priority where one is set, and the dates it runs between. Sort by Starts to see what is next. ## Acting on events [Section titled “Acting on events”](#acting-on-events) 1. **Create.** New event schedules one against an asset and a date. 2. **Send.** Send emails an event out to whoever needs it. 3. **Export.** Export takes the current view out as a file. An event often follows other work: a defect on the [punchlist](/punchlist/) becomes a scheduled repair, and a due [reminder](/reminders/) becomes a booked inspection. # Favorites and pinning > Keep the assets, views, and documents you use most within one click. Keep the assets, views, and documents you use most within one click. ## Favorites in the sidebar [Section titled “Favorites in the sidebar”](#favorites-in-the-sidebar) The Favorites section near the top of the [sidebar](/getting-around/) is your shortlist. Whatever you star, an asset or a saved table view, appears here, so the handful of things you touch every day sit above the full list of assets rather than inside it.  The Favorites section, with a starred asset and a saved document view. ## Favoriting an asset [Section titled “Favoriting an asset”](#favoriting-an-asset) Open an asset and click the star next to its name in the header. The asset joins your Favorites straight away; click the star again to take it off. Nothing about the asset changes: the star only controls whether it shows in your shortlist.  The star beside an asset name. Filled means it is in your Favorites. ## Saving a view to favorites [Section titled “Saving a view to favorites”](#saving-a-view-to-favorites) A saved view is a table cut you have named and kept: a set of folders, [filters](/filtering/), and columns on the [Documents page](/documents/) or any other table. Star a saved view and it lands in Favorites alongside your assets, so a shared or often-used cut is one click from anywhere. ## Pinning documents [Section titled “Pinning documents”](#pinning-documents) Pinning is favoriting for documents, and it is scoped to an asset. Pin the documents that matter on a given asset (the current contract, the latest inspection) and they gather under Pinned documents on that asset and under the Pinned filter on the [Documents page](https://aevy.app/dashboard/documents). Unpinning asks you to confirm, and you can always pin again later.  The pinned view: the documents you have pinned, gathered in one place. # Filtering tables > Narrow any list to exactly what you need, then save the cut or share it as a link. Narrow any list to exactly what you need, then save the cut or share it as a link. ## One filter bar, everywhere [Section titled “One filter bar, everywhere”](#one-filter-bar-everywhere) [Documents](/documents/), [Reminders](/reminders/), [Punchlist](/punchlist/), [Events](/events/), [Contacts](https://aevy.app/dashboard/contacts), [Systems](/systems/): every table in Aevy shares the same filter bar. Learn it once and it works the same on all of them. A filter is a rule on one column, and you can stack as many as you need.  Add Filter lists every column you can filter on, grouped by area. Search narrows the list. ## Adding a filter [Section titled “Adding a filter”](#adding-a-filter) Click Add Filter, then pick the column to filter on. A filter appears as a chip in the bar, ready for you to set its value. Add another for a second rule. ## The three parts of a filter [Section titled “The three parts of a filter”](#the-three-parts-of-a-filter) Every filter chip reads left to right as **field**, **operator**, **value**, and each part is its own click: 1. **Field.** The column the rule applies to, for example Created At. 2. **Operator.** How to compare, for example is after. 3. **Value.** What to compare against, for example 24 Jun 2026. Click any part to change it. The small cross at the end of the chip removes that filter and leaves the rest in place.  A filter reading Created At, then is after, then 24 Jun 2026. Each segment opens on its own. ## Operators follow the field [Section titled “Operators follow the field”](#operators-follow-the-field) The operators you see depend on what kind of column it is, so the choices always make sense for the data: 1. **Text** (a name, a description): contains, does not contain, starts with, ends with, is empty, is not empty. 2. **Dates**: is, is before, is after, is on or before, is on or after, is between. 3. **Options** (a status, a type): is, is not, is any of, is none of.  Picking the operator for a date field. The list changes to fit the column type. ## Combining and clearing [Section titled “Combining and clearing”](#combining-and-clearing) Filters combine with and: a row shows only if it matches every filter in the bar. To start over, Clear removes all of them at once. For the most common cuts, the quick-filter chips above the bar apply a ready-made filter in one click. ## Saving and sharing a view [Section titled “Saving and sharing a view”](#saving-and-sharing-a-view) Your filters, along with the sort, the columns, and the search, live in the page address. Copy the link and whoever opens it lands on the same view, already filtered. When a cut is one you will come back to, save it as a view so it sits along the top of the table for next time. # Getting around > The sidebar, the modes, and how Aevy is organized. The sidebar, the modes, and how Aevy is organized. Vocabulary If a word here is new to you (asset, system, suggestion), the [core concepts guide](/core-concepts/) defines them all. ## The sidebar [Section titled “The sidebar”](#the-sidebar) The sidebar is the fixed column down the left of every page. It stays put as you move around, and it holds five things, top to bottom. 12345 1. **Organization and mode.** Which company you are in, and which mode you are working in. 2. **The menu.** The main pages: Portfolio, Documents, Reminders, and the rest. 3. **Favorites.** The assets and saved views you have starred, for one-click access. 4. **Assets.** Every asset you can see, in one list. 5. **Your account.** Settings, support, and sign out. ## The menu [Section titled “The menu”](#the-menu) The menu is your way into the main pages. What each one holds: 1. **[Portfolio](https://aevy.app/dashboard).** Your starting view across all of your assets. 2. **[Documents](/documents/).** Every document synced from your sources, searchable and filterable. 3. **[Reminders](/reminders/).** Dated tasks, recurring or one-off, so nothing slips. 4. **[Punchlist](/punchlist/).** The defects and issues to track and close out. 5. **[Events](/events/).** Dated milestones and activities across your assets. 6. **[Compliance](/compliance/).** Requirements and obligations to keep on top of. 7. **[Contacts](https://aevy.app/dashboard/contacts).** The people and companies tied to your assets. 8. **[Suggestions](/suggestions/).** What Aevy proposes from your documents, waiting for your review. The number next to it is how many are pending. ## Modes [Section titled “Modes”](#modes) A mode is the lens you work through, and it shapes the menu to fit the job. The modes are Asset Management, Due Diligence, and Owner. To switch, click the organization name at the top of the sidebar and choose one under Mode. That menu is also where you move between organizations, if you belong to more than one. Note Which modes you can use, and which pages each one shows, depend on how your organization is set up. If a page you expect is missing, check which mode you are in.  The mode menu. The current mode carries a check; the others switch you across with one click. ## The rest of the sidebar [Section titled “The rest of the sidebar”](#the-rest-of-the-sidebar) The Assets section lists every asset you have access to, sorted by name; click one to open it. Favorites sits just above it as a shortlist: star an asset or a saved view and it rises to the top, so your daily few are not buried in the full list. The [favorites and pinning guide](/favorites/) covers how starring works. At the very bottom sits your account: open your name for settings, a link to support, and sign out.  The account menu, at the foot of the sidebar. ## The chat panel [Section titled “The chat panel”](#the-chat-panel) One thing lives outside the sidebar: [Squirrel](/squirrel/), the AI asset manager, opens from the right edge of any page. Ask it a question in plain language and it answers from your own documents and records. Point it at a specific record by typing @ in the composer; the [mentions guide](/mentions/) covers how. # Glossary > Every Aevy term in one line. Every Aevy term in one line, linking the guide that covers it. The [core concepts guide](/core-concepts/) holds the same tour in prose. Filter terms... No term matches. | | | | --------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Asset** | A single renewable energy project (solar, wind, storage, EV charging, or a combination); the home base everything else attaches to. See [getting around](/getting-around/). | | **System** | A major technical part of an asset (a turbine, the substation, an inverter station), holding specs and location. See [systems and components](/systems/). | | **Component** | A part inside a system (a gearbox, an inverter); where inspections and defects actually land. See [systems and components](/systems/). | | **Document** | Any file kept against an asset, arriving by sync (no upload step); read, classified, and tracked through processing. See [documents](/documents/). | | **Classification** | Type, subtype, and scope on every document (for example Contract, Amendment, EPC), plus its signature state. See [documents](/documents/). | | **Extracted fact** | A structured value pulled from a document (a fee, a date, an insured sum), always linked to its source line. See [lineage navigation](/lineage/). | | **Related documents** | Version and relation links between documents: superseded, duplicate, amendment, appendix. See [documents](/documents/). | | **Pinned document** | A document kept one click away on its asset and under the Pinned filter. See [favorites and pinning](/favorites/). | | **Suggestion** | A proposed change Aevy drafts from a document; nothing is written until you accept it. See [suggestions](/suggestions/). | | **Confidence** | Aevy’s estimate of how well the document supports a suggestion; a prompt for attention, not a verdict. See [suggestions](/suggestions/). | | **Reminder** | A dated task on an asset: a calibration to run, a certificate to renew. See [reminders](/reminders/). | | **Punchlist item** | A defect or issue on an asset, tied to its system or component, with severity and a status from report to resolution. See [punchlist](/punchlist/). | | **Event** | Something happening on an asset on a date, with a start, an end, and a status. See [events](/events/). | | **Obligation** | Something a contract requires, tracked on the contract’s cadence; unmet means non-compliance. Beta; some screens say “requirements”. See [compliance](/compliance/). | | **Contact** | A person tied to an asset or company, with the details to reach them. See [core concepts](/core-concepts/). | | **Company** | An organization involved with an asset; groups its contacts. See [core concepts](/core-concepts/). | | **Comment** | A note on a record, in the same feed as its history; takes mentions. See [activity log and comments](/activity/). | | **Mention** | Type @ to name a specific record, teammate, or Squirrel in a sentence. See [mentions](/mentions/). | | **Squirrel** | The AI asset manager in the chat panel; answers from your data, cites sources, proposes work as suggestions. See [Squirrel](/squirrel/). | | **Activity** | The audit trail on every record: who did what, when, including Aevy’s own actions. See [activity log and comments](/activity/). | | **Saved view** | A named table cut (folders, filters, columns) kept along the top of a table; starrable into Favorites. See [filtering tables](/filtering/). | | **Mode** | The lens you work through (Asset Management, Due Diligence, Owner); shapes the sidebar menu. See [getting around](/getting-around/). | # Lineage navigation > Trace any fact Aevy pulled from a document back to the exact line it came from. Trace any fact Aevy pulled from a document back to the exact line it came from. ## What this is [Section titled “What this is”](#what-this-is) When a document arrives, Aevy reads it and pulls out the facts it holds: the parties, the dates, the fee, the obligations. You see these on the [document page](/documents/), in the Details panel. Every fact keeps a link to where it came from, so you never have to take an extracted value on trust. Click it, and Aevy reveals the exact source text in the document, highlighted. You confirm a number in the contract itself. ## The document page [Section titled “The document page”](#the-document-page) Open a document and you get two halves: the document itself on the right, and a set of panels on the left. The panels are Work, Requirements, Related documents, and Details, and they are independent, so you can have more than one open at once. The Details panel is where the extracted facts live, grouped by area (classification, parties, term, offtake, fee, and so on). Use Search details to jump to a fact by name.  The document page. Extracted facts on the left, the source document on the right. ## Revealing a fact’s source [Section titled “Revealing a fact’s source”](#revealing-a-facts-source) Click an extracted value and the document scrolls to its page and highlights the source. Take the Sunny Solar power purchase agreement: the **Base fee** reads “€65.00/MWh for Contract Years 1-10”. Click it, and the document jumps to page 3, to the Contract Price section, and highlights **Contract Price: € 65.00 /MWh**. The value you clicked and the line it came from are the same.  Clicking the base fee reveals and highlights its source in the document, on the page it was found. ## The fact menu [Section titled “The fact menu”](#the-fact-menu) Each fact carries a small menu. From it you can: 1. **Reveal in document.** The same jump-and-highlight as clicking the value. 2. **Copy.** Copy the value or its source text. 3. **Accurate or Inaccurate.** Tell Aevy whether it read the fact correctly. This feedback improves later extractions. 4. **Add comment.** Leave a note on the fact for your team.  Every extracted fact can be revealed, copied, marked accurate or not, and commented on. ## From wherever a fact appears [Section titled “From wherever a fact appears”](#from-wherever-a-fact-appears) Facts are clickable wherever you meet them, not only on their own document. In a table, or in an answer from [Squirrel](/squirrel/), the same click opens the document to the same highlighted line. When you are already on the document it reveals in place; from elsewhere it opens the document in a preview and takes you straight to the source. Note on matching Aevy finds the source by matching the text, so short, distinctive values like a fee or a date land on the exact words. A long, clause-length fact lands on the right passage, which may read slightly differently from the extracted wording if the original was laid out across lines. The page is always right; the highlight is as exact as the text allows. # Mentions > Point Aevy at exactly the asset, document, or task you mean, by typing a single character. Point Aevy at exactly the asset, document, or task you mean, by typing a single character. ## What a mention is [Section titled “What a mention is”](#what-a-mention-is) A mention is a way to name a specific record inside a sentence. Type @, pick the thing you mean, and it turns into a labelled chip that points at that exact asset, document, or task. There is no guessing which “PPA” or which turbine you meant: the mention is the record itself.  Type @ and start typing. The picker groups matches by kind and narrows as you go. ## What you can mention [Section titled “What you can mention”](#what-you-can-mention) Ten kinds of record are mentionable. Type @ to see them all, or keep typing to filter to one: 1. **Assets**, **Systems**, and **Documents**: the things you manage and the files behind them. 2. **Tasks**, **Findings**, and **Events**: the work, the issues, and the dated activities. 3. **Contacts**, **Companies**, and **Users**: the people and organizations involved. 4. **Suggestions**: a specific proposal Aevy has made. ## Where you can mention [Section titled “Where you can mention”](#where-you-can-mention) Mentions work anywhere you write in Aevy: in the [Squirrel chat](/squirrel/), in a [comment](/activity/) on any record, and in the notes on a task, a finding, an event, or a contact. Learn the one gesture and it works across all of them. ## What a mention does [Section titled “What a mention does”](#what-a-mention-does) What the mention does depends on where you use it: 1. **In the Squirrel chat**, a mention scopes your question to that record, so “summarize @Sunny Land PPA” tells Squirrel exactly which document to read. 2. **In a comment**, a mention references the record for everyone reading. Mention a teammate to bring them in, or mention Squirrel to ask it to act on the record. 3. **In Squirrel’s replies**, the mentions it returns are live links: click one to open the asset, document, or task.  Squirrel answers with the same references you can mention, as links back to the records and documents. One quirk to know A mention you type is a live, clickable link only in Squirrel’s replies. Where you type it, in the composer, a comment, or a note, it stays as a label that names the record rather than a link you click through. So do not expect the mention in your own comment to be clickable; it is doing its job by naming the record precisely. # Punchlist > Track defects and issues found on your assets, from report to resolution. Track defects and issues found on your assets, from report to resolution. ## What the punchlist is [Section titled “What the punchlist is”](#what-the-punchlist-is) The [punchlist](https://aevy.app/dashboard/findings) is where defects and issues live: a worn brake pad, an expired certification, a rust spot flagged in a statutory inspection. Each item names the asset and the system it sits on, how serious it is, and where it stands, so the open work on your portfolio is one list rather than a scatter of inspection reports.  The Punchlist page. Each row is a defect on an asset, with its status, severity, and system. ## Finding an item [Section titled “Finding an item”](#finding-an-item) The list filters like every table in Aevy (see the [table filtering guide](/filtering/)). What is specific here: 1. **Saved views.** All punchlist, My punchlist, High severity, Blocked, and In progress are prebuilt cuts, each with its own count. 2. **Search.** Search punchlist matches on the item title. 3. **Filter.** Filter and Add Filter narrow by any column, for example severity or status. Each row names the asset, the defect in a line, its stage from Reported onward, and its severity, from Low through Medium and High to Urgent. The system column ties the item to the [system or component](/systems/) it affects, so a defect rolls up to its equipment; once an item is closed, the resolved column carries the date. ## Acting on the punchlist [Section titled “Acting on the punchlist”](#acting-on-the-punchlist) 1. **Create.** New punchlist item logs a defect against an asset. 2. **Send.** Send emails an item out to whoever needs to act on it. 3. **Export.** Export takes the current view out as a file. Defects that Aevy reads from an inspection report arrive as [suggestions](/suggestions/): you review them, and accepting one adds the item here. # Reminders > Track everything due on your assets, from inspections to renewals, in one list. Track everything due on your assets, from inspections to renewals, in one list. ## What a reminder is [Section titled “What a reminder is”](#what-a-reminder-is) A [reminder](https://aevy.app/dashboard/tasks/reminders) is a dated task on an asset: a calibration to run, an inspection to schedule, a certificate to renew. The [Reminders page](https://aevy.app/dashboard/tasks/reminders) gathers them across your whole portfolio, so instead of chasing dates in separate documents you work from one list of what is due and what is overdue.  The Reminders page. Each row is a reminder on an asset, with its category, due date, and status. ## Finding a reminder [Section titled “Finding a reminder”](#finding-a-reminder) The list works the same way as every other table in Aevy, so the [table filtering guide](/filtering/) covers the mechanics. The tools that matter here: 1. **Saved views.** The tabs along the top are prebuilt cuts: All reminders, My reminders, Overdue, Due in 7 days, and High priority. Each keeps its own filters and columns, and the count next to it tells you how many match. 2. **Search.** Search reminders matches on the title and description. 3. **Filter.** Filter and Add Filter build a cut on any column, for example a single asset or a due date range. Each row names the asset, what needs doing, its category, its due date, and its status from To do onward. Sort by the due date column to bring the next deadlines to the top. ## Acting on reminders [Section titled “Acting on reminders”](#acting-on-reminders) 1. **Create.** New reminder adds one, against an asset and a due date. 2. **Send.** Send emails a reminder out, so a deadline reaches whoever needs to act on it. 3. **Export.** Export takes the current view out as a file. Reminders that Aevy proposes from your documents arrive as [suggestions](/suggestions/) first: you review them, and accepting one adds the reminder here. # Squirrel > Your AI asset manager: ask in plain language, and it works the platform the way you do. Your AI asset manager: ask in plain language, and it works the platform the way you do. ## What Squirrel is [Section titled “What Squirrel is”](#what-squirrel-is) Squirrel is the assistant behind the AI Chat panel. You ask it a question in plain language and it answers from your own data: your documents, your assets, your tasks. It answers, and it acts. Squirrel uses the same tools you have, reading documents, searching, looking up facts, drafting records, so when it answers it is answering from the platform.  The AI Chat panel. Ask anything, start from a suggested prompt, or pick up a recent chat. ## Asking a question [Section titled “Asking a question”](#asking-a-question) Open the panel from the right edge of any page. Type your question, or start from one of the prompts under “Try Super Squirrel for”. To point Squirrel at a specific thing, type @ and mention it: an asset, a document, a task. The [mentions guide](/mentions/) covers this. Your chats are kept, so you can return to a thread from “Recent chats” and carry on. ## It uses the same tools you do [Section titled “It uses the same tools you do”](#it-uses-the-same-tools-you-do) Squirrel is not a separate system bolted on the side. It works the platform through the same actions you have, which is why its answers line up with what you see. It can: 1. **Find and read.** Search and summarize assets, systems, and components; find, read, and search across your documents. 2. **Look things up.** Pull extracted facts from documents, and find tasks, reminders, recurring obligations, findings, and events. 3. **Draft, for your confirmation.** Pre-fill a new finding, event, or contact for you to review, and propose [suggestions](/suggestions/) you accept, dismiss, or ask it to revise. 4. **Act in place.** Post a [comment](/activity/) on a record, and remember your preferences across conversations. 5. **Go deeper.** Hand a harder question to a background helper without making you wait for the reply. ## How it works, step by step [Section titled “How it works, step by step”](#how-it-works-step-by-step) Squirrel is agentic: it plans, takes a step, checks the result, and takes the next one, rather than answering in one shot. You can watch it work. Above each answer it shows its activity steps, so you can see it read documents, look up facts, and pull the numbers together. Then it answers with the evidence in hand, and links back to the documents it used.  A question, the activity steps Squirrel took to answer it, and an answer built from the documents it read. ## Answers you can act on [Section titled “Answers you can act on”](#answers-you-can-act-on) Because Squirrel works from your records, an answer is something you can open and act on. It cites the documents behind a claim, and you click a citation to reveal the exact source (see the [lineage guide](/lineage/)). It returns assets, tasks, and documents as clickable mentions that open the record. And when it proposes work, it comes as a suggestion you accept, dismiss, or send back for a change. Nothing it drafts is committed until you say so: Squirrel proposes, you decide.  A worked analysis: the monthly figures Squirrel found, an assessment, and a recommended follow-up. # Suggestions > Review and act on what Aevy reads from your documents. Review and act on what Aevy reads from your documents. ## What a suggestion is [Section titled “What a suggestion is”](#what-a-suggestion-is) Every asset carries a paper trail: inspection reports, grid connection agreements, O\&M plans, specifications, record drawings. The facts you need (a transformer rating, an inspection deadline, a substation name, a change of owner) are in those [documents](/documents/), but reading each one and copying the facts into your asset register by hand is slow and easy to get wrong. A suggestion is Aevy doing that reading for you. When a document arrives, Aevy reads it and proposes a specific change to your data: a component to add, a property to set, a date to record. Each proposal is one suggestion. A suggestion is always a draft. Nothing in your asset register changes until you accept it. You stay in control of what gets written and what does not.  The Suggestions screen: source documents on the left, the review pane on the right. ## Where suggestions come from [Section titled “Where suggestions come from”](#where-suggestions-come-from) Aevy generates suggestions from the documents already attached to your assets. Documents always arrive through the sync from your source, so none are added by hand: the moment a file lands there, it flows onto its asset and into this loop. The flow is the same every time: 1. A document syncs onto an asset from your source. 2. Aevy reads the document and pulls out the facts it can support with evidence. 3. Each fact becomes a suggestion, grouped under the document it came from. 4. The suggestion waits for you on the Suggestions screen until you accept or dismiss it. What decides which documents get reviewed, and how each one is turned into suggestions, is a set of rules. For now, we set those rules up for you: they are not something you configure yourself yet, and you do not need to. Suggestions simply arrive on this screen for you to review. ## Finding your suggestions [Section titled “Finding your suggestions”](#finding-your-suggestions) Open [Suggestions](https://aevy.app/dashboard/suggestions) from the left sidebar. The number next to it is how many are waiting for review. The screen has two panes: 1. **Left: the review list.** Every pending suggestion, grouped under the document it came from. Each group shows the document title, the asset it belongs to, and how many suggestions it holds. 2. **Right: the review pane.** Select a document group on the left and its suggestions open here in full, with all the supporting detail. At the top, a search box and a [filter bar](/filtering/) help you narrow the list. By default the list is filtered to State is Pending, so you see only what still needs a decision. Use Add Filter to narrow further, or Clear to remove filters.  Search and filter to focus the list. Pending suggestions show by default. ## Anatomy of a suggestion [Section titled “Anatomy of a suggestion”](#anatomy-of-a-suggestion) Open a document group and its suggestions fill the review pane, one card each. Here the Sunny Solar operations and maintenance agreement holds five:  One document group open for review: the list on the left, its suggestion cards on the right. The next figure steps into the first card. Every card is built to answer one question: can I trust this enough to accept it? Everything you need to decide is on the card. 12345 One suggestion card. The numbers map to the parts below. 1. **Type.** What Aevy proposes to create or change: a [task](/reminders/), a [compliance item](/compliance/), a [component](/systems/), an [event](/events/), a property update. The badge names it. 2. **Confidence.** How well the document supports the suggestion, shown as a percentage. More on reading this below. 3. **Description.** The proposed change in plain language, with the clause or figure it rests on. 4. **Proposed record.** A preview of exactly what will be created, with its fields (a due date, an owner, a value) laid out before you commit. 5. **Sources.** The document the suggestion came from and the page it was found on, so you can open the original and check for yourself. Why this matters The named source and page are the point. You never take a suggestion on faith: every one traces back to the document and page that produced it, so you can confirm before you commit. ## Reviewing suggestions [Section titled “Reviewing suggestions”](#reviewing-suggestions) You have three moves. ### Accept [Section titled “Accept”](#accept) Accepting a suggestion is the moment the change is written to your asset register. A new component appears on the asset, a property takes its new value, a date is recorded. Until you accept, none of that has happened. In the review pane, you accept a suggestion with its Create button. In the list, each row has its own accept control, so you can act on a single item without opening it. ### Dismiss [Section titled “Dismiss”](#dismiss) Dismissing discards the suggestion. Nothing is written and the suggestion leaves your pending list. Use it when a suggestion is wrong, duplicated, or not worth recording. In the review pane, this is the Dismiss button. In the list, each row carries its own dismiss control alongside accept. ### Accept all [Section titled “Accept all”](#accept-all) When a document is trustworthy and every suggestion under it looks right, Accept all applies the whole group in one action. The number on the button tells you how many suggestions you are accepting.  Accept a whole document at once, or handle each suggestion on its own. You are never forced into the batch: you can accept some items, dismiss others, and leave the rest for later. The list keeps whatever you have not yet decided. ## What suggestions can change [Section titled “What suggestions can change”](#what-suggestions-can-change) Aevy proposes whatever a document supports, across the record types you already work with: 1. **Tasks and [reminders](/reminders/).** An action a document calls for: request a certificate, serve a notice, review a clause before a deadline. 2. **Compliance items.** An obligation to track: an annual report due, an invoice payable within set terms, a certificate to keep current. 3. **Components and systems.** A transformer, a substation, a battery module the document describes but your register does not yet hold. 4. **Asset properties.** A value on an existing asset: a connection voltage, an OEM, a capacity, an address. 5. **Dates and milestones.** A commercial operation date, an inspection deadline, a delivery milestone drawn from an agreement or a report. One document can yield several kinds at once: the agreement above proposes a new task and a set of compliance items, each with its own confidence and source. ## Reading the confidence score [Section titled “Reading the confidence score”](#reading-the-confidence-score) The percentage on each card is Aevy’s own estimate of how well the document supports the suggestion. As a rough guide: | | | | ----------------- | ---------------------------------------------------------- | | 90 to 100 percent | Stated directly in the document, close to word for word. | | 70 to 89 percent | Strongly implied by the document, though not spelled out. | | 50 to 69 percent | A reasonable inference, with some room for interpretation. | | Below 50 percent | Speculative. Read the reasoning closely before accepting. | The ring around the percentage fills and turns green as confidence rises, so a scan down the list tells you which suggestions are near-certain and which want a closer read before you accept. Treat confidence as a prompt for attention: a high score still deserves a glance at the source, and a lower score is worth accepting when the description holds up and the evidence checks out. The score tells you where to look hardest, not what to decide. ## Why it works this way [Section titled “Why it works this way”](#why-it-works-this-way) Two principles shape the whole feature. **Nothing changes without you.** In asset management, wrong data has a real cost: a missed inspection deadline, a component that does not exist, a compliance record that is false. So Aevy never writes to your register on its own. Every suggestion is a draft, and accepting it is a deliberate step you take. **Every suggestion is auditable.** Because each one names its source document and page, a suggestion is not a black box. You can always open what Aevy read and see why it proposed what it did, which is what lets you trust the ones you accept and confidently dismiss the ones you do not. The result is a register you can stand behind: filled faster than by hand, but with a human decision behind every entry. # Systems and components > The equipment on an asset, organized from the whole site down to the single part. The equipment on an asset, organized from the whole site down to the single part. The examples below come in a wind and a solar version. Pick your technology on any tab and every example on the page follows; the choice is remembered as you move between guides. ## Where systems live [Section titled “Where systems live”](#where-systems-live) Systems belong to an asset, so you reach them by opening an asset and staying on its Overview. A system is a major part of the site, and a component is a part inside a system. The structure exists for roll-up: a [defect](/punchlist/) on a component counts against its system, and every system counts toward the asset. That way maintenance and inspections sit in one hierarchy, not in a flat list of equipment. * Wind On a wind farm, a system is a turbine, the substation, or the balance of plant, and a component is a part inside one: a gearbox, a transformer, a rescue kit.  The systems overview. Each card is a system; the tags inside it are its components, colored by whether they need attention. * Solar On a solar plant, a system is the substation or an inverter station, and a component is a part inside one: an inverter, a PV module, a combiner box.  The systems overview for Sunny Solar. Each card is a system; the tags inside it are its components. ## Reading the overview [Section titled “Reading the overview”](#reading-the-overview) The overview opens on cards, one per system, with the components listed as tags and a [compliance](/compliance/) state in the corner. Above the cards, a defect trend and a breakdown by priority (low, medium, high, critical) show where the work is, and an AI findings count shows what we flagged from document analysis in the last thirty days. * Wind On a wind farm the cards are the turbines, the substation, and the balance of plant, with parts like the gearbox and transformer as tags inside each one. * Solar On a solar plant the cards are the substation and the inverter stations: for Sunny Solar you see sub1 alongside inverter stations like Inv11, Inv17, and Inv25. The tags are the parts inside each one: an inverter, PV modules, a combiner box, DC cabling, the mounting structure and piles. ## The four views [Section titled “The four views”](#the-four-views) Four tabs sit above the systems, each a different way to read the same equipment: 1. **Overview.** The cards, with components as tags. 2. **Systems List.** A table with the specifications for the technology: manufacturer and owner, capacity, and operating status, plus hub height and rotor diameter on a wind farm or AC and DC voltage on a solar plant. 3. **Systems Map.** The systems placed on a map, where they have coordinates. 4. **Statutory Inspections.** Inspections rolled up from component to system. * Wind  Systems List. Columns group into system info, manufacturer and owner, technical specs, and capacity. * Solar  Systems List for Sunny Solar, with the solar specs: capacity and AC and DC voltage. ## Inside a system [Section titled “Inside a system”](#inside-a-system) Open a system to see its full record: its standardized reference, type, status, location, and notes, and beneath that the components it holds. Each component is a step deeper: open one to reach its defects, so you move from the site to the system to the part to the specific issue without losing your place. * Wind  A turbine opened from the list. Its components sit at the bottom; a marker shows any that have open defects. * Solar For an inverter station like Inv11 the record reads plainly, a Sungrow SG250HX central inverter at 250 kW AC and 1500 V DC input, in operation, with its inverter, modules, and combiner box listed below.  An inverter station opened from the list, with its make and rating and the components beneath it. ## Adding a system [Section titled “Adding a system”](#adding-a-system) Use New system on the systems view to create one and fill in its reference, type, and status. Components are mapped to systems when your asset is set up, so the parts already sit under the right system when you arrive; the views above are how you read and track them, not where you wire them together.