Changelog

A pass over what happens after you find a funder — so the judgment, the reasoning behind it, and the document you read later all say the same thing.

Assess a whole search in one pass — Once a funder search has produced its briefs, ask Grantable to assess the prospects and it scores every unassessed one in a single run, so the Fit column arrives filled in rather than blank. It shows you first what the pass would cover and waits for you to say go; rows you already scored and rows you dismissed are left alone.

The assessment lands in the brief — Assessing a funder now writes a Fit Assessment section into that funder’s Prospect Brief: the verdict, and the dimensions it was built from. The verdict in your table and the document behind it can no longer drift apart, and re-assessing updates that section rather than adding a second one.

The Fit verdict is one word — Strong, Fair, Weak, or Blocked, with nothing appended. The dimension name that used to trail the verdict meant two different things depending on which verdict it followed, which made it hard to read as either. It now lives on the hover tooltip, where it reads as the explanation it always was.

Blocked says whether there was anything to go on — A funder that failed an eligibility, geography, or gift-size check reads Blocked and offers you a re-check. A funder where the material simply never spoke to that check also reads Blocked, but points you at the brief instead — because asking again would return the same silence. Two different findings, two different next moves.

One judgment per prospect row — The separate data-completeness chip has come off the row. How well a funder fits your organization and how much Grantable knows about that funder are different questions, and a row that answered both at once was answering neither clearly. The row now carries the fit verdict alone.

Long messages stay put in the inbox — A tall email now scrolls within the message itself, so opening one no longer pushes the rest of the page out from under you.

A small one, but it followed people around.

Code blocks read clearly everywhere — When Grantable returns a formatted block in a chat response — a table of contents, a data snippet, a piece of structured output — it now renders in the app’s light theme on every machine. If your computer is set to dark mode, those blocks used to arrive in a colour scheme built for a dark background, sitting on a light one. Same block, same colours, same legibility, wherever you’re reading it.

A funder’s record is never complete. This release moves the question of how complete out of a one-word chip and into the place where there’s room to answer it.

The prospect row says one thing — A funder’s row now shows Fit, and nothing competing with it for your attention. How much Grantable knows about that funder used to sit alongside it as a second word to read, in a column that had no space to explain what it meant. It has moved into the Prospect Brief on every surface that showed it: the table in your conversation, the triage card, and both prospect views on your dashboard. On the dashboard list you can still sort by it.

The brief names what’s missing, and what it costs you — Every Prospect Brief now opens with a plain word for how complete the record is — Excellent, Good, Moderate, Limited, or Poor — and then does the part a chip never could. It names which of five things this funder’s record has and which it lacks: a funder record you can open, the kind of funder it is, a grant size, a deadline, and a source to check the rest against. For each gap, it says what would fill it and what leaving it will cost you — a deadline drawn from a 990 filing with no guidelines page behind it can be a year stale, and that’s worth knowing before you write a word.

Completeness is not fit — The brief says so out loud, because the two are easy to run together. A thoroughly documented funder can still be the wrong one for you, and a thin record can belong to exactly the right one. Knowing which you’re looking at is the difference between a gap you should close and a gap you can ignore.

A simpler prospect row, and a brief that carries the reasoning the row used to crowd — plus somewhere to keep your own working notes on a funder.

One badge on a prospect row — The Fit column is now the only badge on the row. The data-completeness chip that sat beside it has come off every prospect table in Grantable — in chat, on the Dashboard list, and on the funnel — because a completeness word next to a fit word, with no room to explain either, asked you to hold two different questions at once while scanning.

How much we know, in the brief that has room to say it — Each Prospect Brief now opens a section called How Much We Know About This Funder. It leads with the level — Excellent, Good, Moderate, Limited or Poor — says plainly that it measures the record and not your match, then names exactly what’s on file, what’s missing, what would fill each gap, and what it costs you to leave it. “No source on file: check the funder’s own guidelines page before you write anything, because this deadline comes from the 990 and can be a year stale” is worth more than a word on a row.

Your Notes and a Decision Trail — Every brief now carries a checklist you own — confirm the cycle with the funder, find a warm connection, check you have capacity to write it by the deadline — that Grantable writes to and reads back never. Underneath it, a Decision Trail records what happened to the prospect and when, oldest first, so a funder you revisit in three months still tells you why you set it aside.

Screening before a brief is written — Every candidate is checked against your organization’s hard constraints — geography, applicant type, program overlap, grant size, and whether the funder accepts unsolicited applications at all — before Grantable spends a brief on it. Funders that fail a hard check are listed under a Screened out note with the reason, so you can see what was set aside and why; borderline ones stay on the list with the concern named in the brief.

Sources you can actually follow — The Sources table now carries one row per site a fact was genuinely drawn from — a news item, a state grants portal, a foundation directory — instead of a fixed three.

Assessing is a button — When a prospecting run adds funders to your list, Grantable now offers to assess them as a single click rather than a sentence to reply to. Click it and the assessment runs; leave it and nothing is spent.

A dismissal explains itself — Dismissing a prospect now states the Fit that was on it, and a Weak verdict names the one dimension that held it back — so the record of why you passed is specific enough to be worth reading later.

One way to say “not assessed” — A funder nobody has assessed reads Not assessed in the same words with the same shape everywhere it appears. The Dashboard views used to render a bare dash, which couldn’t tell “we haven’t looked” apart from “we looked and couldn’t call it.”

Long emails stay in their own box — A tall message in your Workspace Inbox now scrolls inside its frame instead of pushing the rest of the message out of view.

This release is about work that’s already in your workspace — the documents you’ve drafted, the dates you’re tracking, and the funder research sitting in your chats.

Documents revise in place — Ask the AI to update a Word document or spreadsheet it created and the change lands on the existing file. Same file, same folder, same place in your tree, and every revision is captured as a version you can return to.

Version history for Word and Excel files — Version history is now available from the document viewer, not only the text editor. Open any generated .docx or .xlsx, step back through its saved versions, preview one, and restore it.

A confirmation step before deleting — Removing anything from the file tree now asks first, files and folders alike, so a stray click on a menu never costs you a document.

Deadlines read the same in every view — The Overview, the list, the board, the calendar, the sort order, and CSV exports all read a due date the same way now, whatever format it arrived in. A project shows up on the day it’s actually due, in every view, including ones imported or set by the AI.

Funder tables open with the answer — When a funder search finishes, the prospect table opens alongside the results instead of waiting behind a control, and the search hands the AI the next step so results turn into a table you can triage. Close the table and it stays closed — that’s a preference, not a state the AI argues with.

Every table card opens its own table — A conversation can hold several funder tables. Each table’s card now addresses its own table, so clicking the one you’re reading about opens exactly that one.

A pass over the words Grantable uses to describe a funder — so the same funder means the same thing everywhere you see it.

One set of words for Fit — A funder’s fit is now always one of four words: Strong, Fair, Weak, or Blocked. The same word appears on the prospect row, in the Prospect Brief, and in the Grant Opportunity Brief, so a funder can’t read one way in your table and another way in the document it links to. Blocked is reserved for a genuine deal-breaker — a failed eligibility, geography, or gift-size check — and says so plainly rather than burying it in a middling result.

Fit is a judgment, not a number — The 0–100 score and the one-word instruction that used to sit at the top of a Grant Opportunity Brief are gone. In their place, the brief reports the Fit level next to the evidence that produced it: the three scored dimensions, the gaps to close, and the dimension actually holding the funder back. A number told you where a funder landed; the dimension that binds it tells you what to do about it.

Grantable tells you when it hasn’t looked yet — A funder nobody has assessed reads Not assessed yet. “We haven’t evaluated this” is a real state, and it now reads as one — clearly distinct from a funder that was assessed and came back average.

Clearer language for funder status — Funders between grant cycles now read Between cycles, and funders with no open call read No open call — plain descriptions of where a funder is, in place of the internal shorthand that was showing through.

A calmer workspace while the AI writes — A prospecting run can produce a brief for every funder it finds. Every file Grantable writes is now collected in a single receipt in the conversation, naming and linking each one, and the document panel opens when you click a file. Your workspace stays as you arranged it.

This release is about the surfaces you work in every day — the table a funder search hands you, the word “fit” wherever it appears, and the dates behind your pipeline.

Prospect tables open with the results in view — When a funder search finishes, the table now opens on the answer instead of waiting behind a collapsed pill above the composer. Collapse it and it stays collapsed — Grantable treats that as your preference and won’t reopen it behind you. Checklists keep the pill, since a checklist is a progress readout you already know the state of.

Every table card opens its own table — A conversation can hold several prospect tables, and each card in the thread now addresses its own. Click the card from your New Mexico search and you get the New Mexico table, however many searches came after it.

Fit means fit, in chat as well as in the table — Ask Grantable about a folder of prospect briefs and you now get the same reading the table shows: a real verdict where an assessment has run, and a plain not assessed where one hasn’t. Prospect briefs also carry their triage state and pipeline stage, so a question about your shortlist comes back matching what’s on your screen.

One reading of Fit everywhere — The Fit column now names which of the two un-assessed states a funder is in: Not assessed when nobody has judged it yet, Assessed — inconclusive when the assessment ran but the funder material was too thin to call. How complete Grantable’s data on that funder is now sits beside the verdict as its own labelled chip, so completeness never reads as a verdict. Same chip, same colors, on every surface where fit appears.

Deadlines in one shape — The dashboard list, the calendar, deadline sorting, and CSV export all read a due date through one shared rule, so a deadline is the same calendar day wherever you look. Anything that isn’t a readable date — a TBD, a bare year — is treated as undated and shown in the calendar’s undated row, rather than turned into a day nobody meant.

Scheduled runs across Pro and Pro+ — Scheduled chats now run on Pro+ workspaces alongside Pro.

Three improvements to the moments where Grantable has to point at exactly the right thing — the date on a grant, the table you clicked, and the runs you set to repeat.

One rule for every deadline — The dashboard reads a grant’s due date the same way everywhere: on the list, on the calendar, in the sort order, and in your CSV export. Whatever shape the date arrived in — typed by you, pulled from an RFP, or written by the AI — it lands on the calendar day it’s actually due, and a grant nobody has dated yet collects in the No due date row where you can find it.

Every table card opens its own table — A conversation can hold several prospect tables. Each table’s card in the thread is addressed to the table it belongs to, so clicking the card for your New Mexico search opens the New Mexico results, however many searches came after it.

Scheduled runs on Pro and Pro+ — Recurring runs you set up in chat fire on schedule on both paid plans.

A round of precision work on the details you act on — when something is due, how strong a match is, and the work that runs while you’re away.

Deadlines are calendar dates, everywhere — A grant deadline is a day on the calendar, not a moment in time, and Grantable now treats it that way from end to end. The date you see on a prospect card, in the dashboard list, and on the calendar is the same date, and it stays the same date whatever time zone you work in.

Fit language is consistent between chat and the table — When you ask Grantable about a set of funders it has already researched, it now describes each one with the same fit classification you see in the prospect table — and says plainly when a funder hasn’t been assessed yet, rather than reaching for a number that isn’t a fit judgment. Asking about a folder of prospect briefs is also a single look now instead of a file-by-file read, so answers come back faster.

Each table opens from its own card — A conversation can hold several tables. Opening one from its card in the thread now brings up that table, so a chat that has produced more than one slate stays easy to move around in.

Scheduled tasks run on Pro and Pro+ — Recurring scheduled tasks now run on schedule for every workspace on a paid plan. If you had a weekly scan or a morning briefing set up, it will pick up on its next scheduled run.

Three improvements to the moments where your existing work comes into Grantable — from a team drive, from a colleague’s invitation, or from your V1 account.

Shared drives in the Google Drive picker — The Drive picker now opens with two tabs: your own My Drive files and the shared drives your organization has given you access to. Grant material usually lives on a team drive rather than in anyone’s personal Drive, and now you can import from either in the same flow.

Invitations find you when you sign in — If a teammate invited you and you sign in with that email address, Grantable recognizes the waiting invitation and adds you to the workspace — no need to hunt down the original email. Choosing Skip for now is remembered, so you’re only asked once.

Organization profile on V1 import — Importing a V1 workspace now brings the full set of organization details you’d filled in there — mission, focus areas, budget, team, location — into your workspace Profile.md. That file is read in every chat in the workspace, so a richer profile means better first drafts. If your V1 account didn’t have those details, Grantable asks you for them right after you land.

Write and send email without leaving a conversation, plus a more honest read on what Grantable actually knows about a funder.

Email drafts in chat — When a conversation calls for an email, Grantable now stages it as an editable compose card right in the thread — To, Cc, subject, and body, all editable before anything goes out. Send it from your workspace inbox so the exchange stays threaded to the conversation, or hand it off to your own mail client pre-filled. Sends from a workspace carry that workspace’s identity.

Signal strength on prospect tables — A funder’s row now shows how complete Grantable’s data on that funder is — Excellent through Poor — rather than a guess at fit. Signal and fit are different questions, and a funder with excellent data can still be a poor match. A categorical fit verdict appears once a prospect has actually been assessed, so the table tells you what to read next without pretending to have already decided.

Cloud imports into the folder you’re in — Start a Google Drive import from any folder’s own menu and the files land in that folder. The Drive picker is easier to reach, easier to close, and clearer about what to do when a browser extension gets in its way.

Faster funder pages — Related-funder lists across Grantable’s funder directory are now precomputed, so profile pages and their suggested funders load noticeably faster.

Steadier sessions — Uploads, file renames, and workspace loading handle intermittent connections more gracefully.

Files from your cloud drives now land exactly where you want them, and every billing change tells you plainly what happens next.

Import straight into a folder — Open any folder’s ⋯ menu in the file sidebar and choose a provider under Import from. Everything you select arrives inside that folder instead of at the top of your library, so a set of RFPs or a year of board reports stays organized from the moment it lands.

Clearer cloud import menu — The + New menu now groups Google Drive, Dropbox, OneDrive, and Box under an Import from heading, so it reads as a way to add files rather than a list of app names.

Smoother cloud connections — Connecting a drive and picking files is more dependable end to end: the picker opens and closes cleanly, your selection is always carried through, and if a browser extension gets in the way, the message names it so you know what to turn off.

A cancellation flow that tells you what happens — Canceling a plan now ends on a screen stating the exact date your access ends, followed by a confirmation email. Your workspaces and documents are kept, and you can reactivate any time before that date.

A redesigned sidebar that puts your conversations front and center, plus tighter privacy controls for cloud integrations.

Chats tab in the sidebar — Your conversation history now lives in a dedicated Chats tab in the left sidebar. Browse, search, and jump back into past conversations without opening the full search overlay. The sidebar navigation is also collapsible, giving you more room to focus when you need it.

Scoped Google Drive access — When you connect Google Drive, Grantable now uses a file picker that only requests access to the specific files you select — not your entire Drive. This means fewer permission prompts and a clearer privacy boundary: Grantable only sees what you explicitly share.

Referral credits — Refer a colleague and both of you receive account credit. The referral program now rewards both the inviter and the new user.

Chat reliability — Conversations are more resilient through long working sessions, with smoother recovery from intermittent network blips.

A dedicated inbox for receiving emails inside Grantable, and a faster path from a funder page on the web into an AI assessment of that funder.

Workspace Inbox (Pro) — Receive emails directly in your Grantable workspace. Incoming messages are analyzed by the AI, which can draft replies using your organization’s context and Grantable’s help documentation. Save attachments straight to your file library with one click. Available on Pro and Pro+ plans.

Deep-link funder assessment — Links from Grantable’s marketing pages and funder directories can now drop you directly into a prospecting conversation with that funder pre-loaded. Sign in (or sign up) and land right in the assessment — no copy-pasting funder names.

A new top-tier plan, smarter onboarding prompts, and branded communications.

Pro+ plan tier — A new top-tier plan for larger teams, with room for up to 10 members and priority support. A new upgrade dialog shows exactly which capabilities unlock at each tier, so you can compare without leaving your workflow.

Context-aware quick actions — The quick-start suggestions on the home screen now adapt to your workspace. A new account sees prompts like “Create my org profile” and “Upload existing materials,” while an active workspace surfaces suggestions tied to your current grants, recent conversations, and upcoming deadlines. Same chat, smarter starting point.

Branded email templates — Magic links, team invitations, and other transactional emails now arrive with Grantable branding and clean formatting instead of plain text.

Final hardening ahead of wider availability — multi-workspace billing, plan-aware features, and in-app support.

Multi-workspace billing — Organizations with multiple workspaces now get a single unified bill. Payment method defaults, cancellation, and account scoping all work correctly across workspaces.

Plan-aware features — Prospecting tools and scheduled tasks now show clear indicators of which capabilities are available on Pro.

Interactive tables — The AI can present structured data as interactive tables with checkboxes, sorting, and a detail panel — useful for reviewing funder shortlists or comparing grant opportunities side by side.

In-app support — A contact support button now appears on error pages, and the support modal links directly to help documentation for self-service troubleshooting.

Login reliability — Improved magic link compatibility with enterprise email security tools that pre-fetch URLs.

A focused round of reliability improvements across chat, notifications, and cloud integrations.

Chat reliability — Improved chat stability under intermittent connections, with more consistent message delivery and stream recovery.

Notifications & comments — Comment notifications, @mention rendering, and email reply formatting now work reliably across all notification channels.

Cloud integrations — Google Drive and OneDrive connections now handle token refresh automatically, and file imports surface clear errors when intervention is needed.

Inline dashboard editing — List view items on the dashboard can now be edited inline with click-to-edit and keyboard navigation between rows.

Help docs in chat — Ask the AI a “how do I…” question and it can now search Grantable’s help documentation to give you accurate, up-to-date answers about the platform.

A unified navigation system, automated task scheduling, a guided V1 migration, and broader sign-in options.

Unified tabs — Files, chats, search results, and the dashboard now all open as tabs in a single view. Switch between a draft proposal, a funder chat, and your dashboard without losing context.

Scheduled tasks — Set any AI workflow to run on a recurring schedule — daily funder monitoring, weekly prospect scans, or monthly report generation. Review results and manage schedules from the notifications panel. Available on Pro.

V1 migration — Existing Grantable V1 users can migrate their workspace, files, and settings to V3 with a guided migration flow.

Microsoft sign-in — Sign in with a Microsoft account in addition to Google and email magic links.

A redesigned starting experience, team file sharing, and a refreshed visual style.

New home page — The home screen now features contextual suggestion chips based on your workspace content, making it easy to jump into common tasks — start a prospect scan, draft a follow-up, or review a budget — without typing a prompt from scratch.

File sharing — Share any workspace document with teammates via internal links, public links, or email. Recipients can view and comment without needing their own Grantable account.

AI edit notifications — When the AI modifies a document you have open, a toast notification appears with a collapsible diff showing exactly what changed.

Interactive prospect list — The AI can present funder prospects as an interactive list where you can dismiss, star, or drill into any result directly from the chat.

Tabbed document viewer — The right panel now supports browser-like tabs, so you can keep multiple documents open and switch between them while chatting.

Visual refresh — Updated typography with DM Sans, softer colors, and a more compact editor toolbar for a cleaner workspace.

Cloud integrations, AI-generated documents, and a more resilient infrastructure.

Cloud storage imports — Connect Google Drive, OneDrive, Dropbox, or Box to import files directly into your workspace. Bring in past proposals, budgets, and org documents without manual uploading.

AI document generation — Ask the AI to produce a Word document or spreadsheet and it generates a real .docx or .xlsx file — not just text. Download it or keep editing in the workspace.

Live editor streaming — When the AI drafts or revises a document, changes stream directly into the editor in real time. Watch a letter of inquiry or budget narrative take shape as the AI writes.

Model tiers — Choose between Auto (balanced), Pro (most capable), and Fast (quick responses) depending on the task. Auto selects the best model for each request automatically.

Stream resilience — Chat sessions now persist across page navigation and network changes. If anything interrupts, the conversation picks up right where it left off.

Major upgrades to search, the chat input, and how documents render in the workspace.

Unified search — A single search bar now searches across files, folders, and chat history simultaneously. Results are grouped by type with keyboard navigation.

@ mentions and / commands — The chat input now supports rich text. Type @ to reference a file or folder as context, or / to invoke a skill like Grant LOI or Prospect Research — with autocomplete as you type.

Native document viewers — Word documents, spreadsheets, and CSVs now render natively in the right panel instead of being converted to PDF. Faster loading, better fidelity, and you can see your documents exactly as they are.

Version history — Documents now track version history. Rename, delete, or restore previous versions from the editor header menu.

Resizable sidebar — Sidebar sections (Files, Chats, Skills) can be individually resized and collapsed, so you can focus on what matters for the task at hand.

A full round of polish to the chat experience and how files surface in conversations.

File action cards — When the AI creates or edits a document, it now appears as a clickable card inline in the chat — not just a text mention. One click opens it in the viewer.

Clickable file paths — Any file path the AI references in a response is now a live link. Click to jump straight to that document.

Activity indicator — While the AI is working, a status line shows what it’s doing (searching funders, drafting a narrative, analyzing an RFP) instead of a generic spinner.

Web search results — When the AI searches the web for funder or program information, results appear as structured cards with titles, snippets, and source links.

Skills — Reusable prompt templates for common grant workflows (LOI drafts, prospect research, budget narratives) can now be invoked in any conversation, giving the AI specialized instructions for each task.

Grantable V3 is a ground-up rebuild of the platform around a single AI chat interface. Grant writing, funder research, document management, and prospecting now live in one workspace powered by conversation.

Chat-first workspace — The home screen is now a prompt input. Describe what you need — a letter of inquiry, a funder landscape scan, a budget narrative — and Grantable gets to work across your files and funder data.

Document upload onboarding — New users can upload a ZIP archive of their existing grant materials during setup. Grantable processes everything into an organized file library and generates an organization profile automatically.

Live file tree — The sidebar updates in real time as the AI creates, moves, or renames documents during a conversation. No refresh needed.

Editor-to-chat context — Select any text in a document and send it to chat with one click. The AI sees exactly what you’re looking at, so you can ask follow-up questions about specific sections.

Transparent tool use — Every action the AI takes now shows a human-readable label explaining what it’s doing and why, so you always know what’s happening.