Claude tutorials, resources, and Skillsbuilt for the work you actually do.
A curated library for two communities we know best: law firms and nonprofits. Buy a single Skill, follow a tutorial series, or download a resource your team can use today.
Tutorials
Step-by-step walkthroughs for real day-to-day workflows.
Resources
Prompt packs, templates, and playbooks you can use today.
Skills
Drop-in Claude Skills that plug into your existing tools.
Generic AI isn't built for law firms.
Most attorneys using AI spend more time fixing output than they would have spent drafting it themselves. The AI doesn't know who you are, how you write, or which client you're working on. So every session starts cold. Every draft needs rewriting. And the time savings you were promised disappear into cleanup.
That's not an AI problem. It's a context problem.
It forgets your clients
Every new session is a blank slate. You re-explain the matter background, rebuild the client context, restate the firm's approach. Every. Single. Time. Not because the AI is bad — because it has nothing to work from.
It doesn't sound like you
Generic output means constant rewriting. The AI doesn't know your firm's voice, your writing standards, or the specific language you use with clients. What comes back is technically correct and completely off-brand.
Nothing stays organized
No matter memory. No file structure. No continuity between sessions. Just a chat window that starts fresh every time you open it, with no record of what was decided, drafted, or agreed last week.
CoWork OS gives Claude the context it's missing.
A structured folder system that loads your firm's identity, matter history, and writing rules before Claude does anything. The result is an AI that sounds like your firm, remembers your clients, and produces output you can actually use.
Persistent matter memory
Every client gets a folder with a brief, a memory log, and an outputs directory. Claude reads the whole thing before it touches your work. When you pick up a matter next week, it picks up exactly where you left off.
Your voice, every time
Your firm's voice profile and legal writing rules travel with every session. Banned phrases. Tone standards. Communication register. What comes back sounds like it was written by your firm — because it was briefed by your firm.
Confidentiality built in
Client matters never cross. CoWork OS is hardwired to flag anything that could mix client information across matters. Structural confidentiality — not just a policy.
Everything in the box.
The folder structure
ABOUT ME/
Six identity files Claude reads every session without exception. Your firm profile, communication voice, legal writing rules, tool map, specialist routing, and session memory. This is what makes the AI sound like your firm rather than a template.
WORK AREAS/
Five organized work areas covering every type of firm work: Client Matters, Practice Development, Marketing, Operations, and Admin. Each matter gets its own brief, memory log, and outputs folder.
RESOURCES/
Templates, guides, plugins, and skills. Everything you need to work immediately.
CLAUDE.md
The operating brain. Global instructions that govern how Claude behaves across every session, every matter, every document.
10 legal templates
Every template is pre-loaded with your firm's voice profile and writing rules. Claude uses the structure — you don't copy the content.
- Matter brief
- Client intake
- Engagement letter
- Demand letter
- Client status update — three formats (substantive, nothing-new, post-meeting)
- Billing narratives
- Legal research memo
- Email formats — client, opposing counsel, referral, cold outreach
- Weekly review
- Matter memory log
Up and running in 15 minutes.
Download the zip
One file. Everything included.
Unzip and place
Extract the folder and put it wherever you keep your work. Create a separate folder for each firm you install it for — CoWork keeps them completely isolated.
Load into CoWork
Open the CoWork desktop app and point it at the folder. That's the connection.
Trigger first run
Open a new chat. Claude detects the setup file automatically and walks you through onboarding — no special commands needed.
Fill in your firm
Answer questions about your practice, your voice, your team, and your tools. Takes 10–15 minutes. Once it's done, Claude knows your firm.
Each law firm client gets their own separate folder. Claude stays completely separated between firms — confidentiality is structural, not just promised.
Built for attorneys who are done fixing AI output.
Solo practitioners
Handle the administrative load of a larger firm without the overhead. Engagement letters drafted and ready to review. Billing narratives from rough notes in seconds. Client updates written in your voice before you've typed a word. You practice law — CoWork OS handles the paper.
Small and mid-size firms
Consistent firm voice across every attorney. Every matter organized the same way. New associates and paralegals get up to speed in one session because the brief and memory log are already there. The firm runs like a firm, not like five individual practitioners.
Legal ops and office managers
A system for the whole firm, not just the power users. Templates and workflows that every attorney can follow. Matter organization that doesn't depend on individual habits. One standard — whoever opens the file next knows exactly where they are.
One price. Your firm keeps it.
Law Firm CoWork OS
$149
one-time
- Complete folder structure and CLAUDE.md operating file
- 6 identity and memory files — pre-built, ready to fill in
- 10 legal document and workflow templates
- Installation guide (HTML — opens in any browser)
- How-to-use guide (HTML — reference anytime)
- Personal assistant plugin
- Specialist sub-agent builder plugin
- Lifetime access — no subscription, no renewal
- Install for every attorney in your firm
Instant download. Works with Claude and the CoWork desktop app. One folder per firm — install for as many attorneys as you have. All templates require attorney review before use.
Frequently asked questions
Generic AI doesn't understand nonprofit work.
Development directors and EDs using generic AI tools run into the same wall: the output is fine until it isn't. It doesn't know your theory of change. It doesn't know how you talk about the people you serve. It doesn't know what your last grant report said or what this funder cares about. So you rewrite everything. The time you thought you were saving goes into cleanup — and the pressure of the next deadline is already building.
It doesn't know your mission
Generic AI produces generic mission language — vague, jargon-heavy, interchangeable with any other nonprofit in your sector. Funders read hundreds of proposals. Program officers notice immediately when the narrative could have come from anyone. Your org's specific theory of change, your community's specific need, your program's specific approach — the AI doesn't know any of it.
It invents impact claims
Ask generic AI to write a needs statement or an impact narrative without providing data, and it will. It will produce plausible-sounding statistics, confident outcome claims, and compelling-but-unverifiable impact language. Funders verify. The wrong number in a proposal damages a relationship that took years to build.
Every session starts from scratch
No funder memory. No program history. No record of what the last proposal said, what the program officer responded, or what was promised in the last grant report. You rebuild context every time you open a chat window — and the AI has no way to know what it's supposed to be consistent with.
CoWork OS gives Claude the context it's missing.
A structured folder system built specifically for nonprofit operations. It loads your organization's identity, funder relationships, program data, and communication standards before Claude writes a word. The result is AI that sounds like your org, respects your data, and produces work you can actually submit.
Funder memory that persists
Every funder relationship gets its own folder — brief, memory log, past proposals, report notes, program officer context. Claude reads it before every session. When you return to a funder relationship next quarter, everything is there: what you proposed, what was funded, what was promised, what the relationship history looks like.
Four communication registers, built in
Nonprofits don't write in one voice — they write in four. Funder-facing copy is precise and evidence-based. Donor-facing copy is warm and story-driven. Constituent-facing communications are plain-language and accessible. Internal documents are direct and practical. CoWork OS knows the difference and applies the right register automatically.
The data-first rule
CoWork OS will not draft impact claims, outcome statements, or needs statement content without supporting data. If the data isn't there, Claude asks for it before writing. This isn't a limitation — it's protection. Your credibility with funders and donors depends on claims you can defend. CoWork OS keeps that standard enforced, session after session.
The feature that pays for itself the first time you use it.
Grant writing is the highest-stakes, most time-consuming writing work in a development office. A strong proposal takes days. A weak one — or a late one — closes a funder relationship that took years to build. CoWork OS was designed around this work first.
The grant proposal scaffold
Our flagship template covers every standard section funders require: organization background, statement of need, program description, evaluation plan, sustainability plan, and budget narrative. Each section has built-in prompts that ask for the data before Claude drafts. No invented statistics. No vague impact language. A complete, data-grounded first draft ready for your review.
The grant report template
Honest reporting builds funder relationships that last. Our report template is structured around the real reporting cycle — comparing actuals to targets, explaining variances with context and learning, and setting up the renewal conversation. Funders don't need everything to have gone perfectly. They need to trust you.
The LOI template
A two-page pitch format that answers the five questions every program officer needs answered: who you are, what you want to do, why it matters, why you can do it, and what you're asking for. Mirror the funder's language. Lead with the need. State the amount early.
Funder relationship memory
Every funder gets a folder. The brief captures their priorities, relationship history, and prior application notes. The memory log tracks what was submitted, what feedback was received, and what the program officer said at the last site visit. The next proposal builds on everything that came before it — not from a blank page.
Everything in the box.
The folder structure
ABOUT ME/
Six identity files Claude reads every session without exception. Your org profile, four-register voice system, dignity-first writing rules, tool map, specialist routing, and session memory. This is what makes Claude sound like your org rather than a sector template.
WORK AREAS/
Six organized areas covering every type of nonprofit work: Grants-Funding (the flagship), Donor-Relations, Communications, Programs, Operations, and Admin. Every grant, campaign, and initiative gets its own brief, memory log, and outputs folder.
RESOURCES/
Templates, HTML guides, plugins, and skills. Everything you need to work immediately.
CLAUDE.md
The operating brain. Governs every session, enforces the data-first rule, applies register awareness automatically, and flags anything that could compromise impact integrity or donor confidentiality.
12 nonprofit templates
Every template is pre-loaded with your org's voice profile and the data-first rule. Claude uses the structure — you bring the data.
- Grant proposal scaffold — data-first, section by section
- Grant report — interim and final, honest-reporting format
- Letter of inquiry (LOI)
- Donor acknowledgment — first gift and major gift formats
- Appeal letter — with ethical guardrails built in
- Donor impact update
- Board report
- Impact summary / annual report section
- Newsletter — three-section format
- Program documentation brief
- Volunteer and community communication
- Project memory log
Up and running in 20 minutes.
Download the zip
One file. Everything included.
Unzip and place
Extract the folder and put it wherever you keep your work. Each organization you install it for gets its own separate folder.
Load into CoWork
Open the CoWork desktop app and point it at the folder.
Trigger first run
Open a new chat. Claude detects the setup file and walks you through onboarding automatically.
Fill in your organization
Answer questions about your mission, programs, funding mix, and communication voice. The voice profile takes the most time and is worth doing carefully.
Read the data-first rule
Bring your program outcome data and at least two external data points for any needs statement. Claude does the rest.
Each organization gets its own separate folder. Multiple staff can use the same CoWork OS install — one purchase covers your whole team.
Built for the people doing the most writing with the least time.
Development directors and grant writers
The proposal deadline is in ten days. You have three other deadlines behind it. CoWork OS drafts the needs statement from your program data, structures the narrative around the funder's priorities, and flags every impact claim that needs a data source before you read it. You review, refine, and submit — instead of drafting from scratch at 10pm.
Executive directors of small nonprofits
You write the grants, the donor letters, the board reports, the newsletter, and the program documentation — and then you do your actual job. CoWork OS handles the first draft of all of it in your org's voice and at your org's standard. Not generic AI output you spend an hour fixing. Something you can actually use.
Communications and program staff
Whether it's a volunteer recruitment email, a program participant update, or an annual report section, CoWork OS applies the right register automatically. Funder-facing precision for proposals. Warm story-driven language for donors. Plain language for participants. You don't have to think about the register — it's already loaded.
We built something most AI tools won't.
Nonprofit CoWork OS will not generate impact statistics, outcome claims, or community needs data that isn't grounded in real information you provide.
This is deliberate. Funders verify claims. Program officers at major foundations have seen every version of AI-generated impact language — and the ones that don't hold up damage the relationship that generated the application. Donors who discover that impact numbers were inflated don't give again.
When Claude needs data to draft a needs statement, it asks you for it. When an impact claim in a draft doesn't have a source, it flags it before delivering the document. When a proposal section requires specific program outcome data, it waits for you to provide it.
This rule is structural — it's built into the CLAUDE.md file that governs every session. It can't be turned off by accident.
Your credibility with funders and donors is worth protecting. CoWork OS protects it.
One price. Your whole team uses it.
Nonprofit CoWork OS
$49
one-time
- Complete folder structure and CLAUDE.md operating file
- 6 identity and memory files — pre-built, ready to fill in
- 12 nonprofit document and workflow templates
- Grant writing guide — how to use CoWork OS for proposals and reports
- Installation guide (HTML — opens in any browser)
- How-to-use guide (HTML — reference anytime)
- Personal assistant plugin
- Specialist sub-agent builder plugin
- Lifetime access — no subscription, no renewal
- Install for your whole team — one purchase per organization
Instant download. Works with Claude and the CoWork desktop app. One folder per organization — your whole team can use the same install. All grant proposals and donor communications require human review before submission or distribution.
Frequently asked questions
Need something custom for your team?
We build bespoke Claude Skills, internal tutorials, and rollout playbooks for firms and nonprofits ready to operationalize AI.
Talk to us