Grant Assistant and GrantGunner both match organisations to funding and draft proposals with AI. Grant Assistant puts its own trained model at the centre and prices each customer by quote. GrantGunner is a shared workspace with published prices, where your people and AI agents such as Claude, ChatGPT and Cursor work on the same applications.
Facts about Grant Assistant come from grantassistant.ai on 16 October 2026. Its site carries FreeWill branding.
What Grant Assistant does well
- Discover. Semantic matching ranks opportunities by fit with your mission, programmes, geography and impact areas, rather than by keywords.
- Respond. It drafts proposals that it says stay true to your mission and voice, and it says it can do so "in a third of the time".
- Specialist training. It says its AI is "trained on over 7,000 successful grant proposals" and refined by experienced grant writers.
- Experienced team. Its founders include former senior staff from large nonprofits, USAID and UN agencies, and it calls out international development organisations as a key audience.
- Support. Its FAQ mentions live chat with a partner success team during business hours.
- Data handling. It says your data is not used to train general language models.
Pricing is not published. Its FAQ says pricing is "customized based on organization size and grant program scope", and you get a quote after a demo.
What GrantGunner does differently
- Use the AI you already have. Every plan includes an MCP server with over 80 tools for Claude, ChatGPT, Cursor or any MCP client. Your assistant can read your profile and documents, search past answers and save drafts.
- Or use our drafter. You choose the model, tone and language. It skips questions it cannot answer from your facts and leaves a note for a person.
- Scout finds grants that fit your profile and checks eligibility before adding them to your pipeline.
- Activity feed and presence. Every edit records who made it and where, and agents get warned before editing an application someone else has open.
- Owners and stored answers.
- Published prices: Free, Starter £25 a month, Pro £80 a month, Enterprise by contact.
Side by side
| GrantGunner | Grant Assistant | |
|---|---|---|
| Matching | Scout agent with eligibility check | Semantic matching (Discover) |
| Drafting | Built-in drafter or your own AI via MCP | Respond, trained on 7,000+ winning proposals |
| Bring your own AI via MCP | Yes, every plan | Not mentioned |
| Team visibility | Activity feed, presence, owners | Not detailed on pages we read |
| Support | Email; founder support on Enterprise | Live chat with partner success team |
| Pricing | Published; free plan | By quote after demo |
| Focus | UK first; charities, startups, research | Mission-driven orgs, international development |
GrantGunner prices as published on 16 October 2026.
Where Grant Assistant is the better fit
- You work in international development and value a team with USAID and UN experience.
- You want a model trained specifically on winning proposals rather than a general assistant.
- You want a vendor relationship with a partner success team and pricing tailored to your organisation.
Where GrantGunner is the better fit
- You want to know the cost before you talk to sales. Our prices are on the website, and the free plan includes the MCP server.
- Your team already uses Claude or ChatGPT. Your organisation may already pay for a capable model. GrantGunner gives it your real data and a shared place to save work, so you are not paying twice for AI.
- Several people and agents share each bid. The activity feed shows who changed what, and drafter notes tell reviewers where to look.
- You want past answers to build up. Approved answers stay in your workspace and are searchable by both people and agents.
As of October 2026, inviting extra colleagues into a shared team workspace on GrantGunner needs Enterprise. Agents you connect work on every plan.
Checking any "trained on winning proposals" claim
Many tools say their AI learned from successful bids. That can help with structure and tone. It cannot know your organisation's facts. Whatever you choose, test it this way:
- Give it a real question from a live application.
- Check every number and claim in the draft against your own records.
- See what it does when a fact is missing. Does it guess, or does it stop and ask?
GrantGunner's drafter is built to stop. When the facts run out, it skips the question and leaves a note saying what a person needs to confirm.
A worked example
A UK international NGO applies to a foundation's health programme.
On GrantGunner, Scout adds the programme to the pipeline. The grants manager owns it. The programme lead asks Claude, connected over MCP, to draft the theory of change from the project documents in GrantGunner. The drafter answers organisational questions and flags the safeguarding policy date for checking. The director reviews the feed and approves, and the grants manager submits on the foundation's portal.
On Grant Assistant, the same NGO would use Discover to confirm fit and Respond to draft, with support from the partner success team, at a price agreed after a demo.
Questions to ask before you choose
Run the same checks on Grant Assistant and on GrantGunner before you commit:
- Can I see a draft built from my own documents? Sample data in a demo tells you very little about your real bids.
- What happens when the AI does not know an answer? You want it to stop and flag the gap, not fill it with a plausible number.
- Who can see what changed? If two people and an AI agent touch one application, you need a record of who wrote what.
- Can we use the AI tools we already pay for? If your organisation has ChatGPT or Claude licences, connecting them saves money and retraining.
- What will it cost when we grow? Check seat limits, credit or word caps, and any success fees.
Moving your work into GrantGunner
If you try GrantGunner after using Grant Assistant or anything else, setup takes an afternoon, not a project:
- Profile. Give GrantGunner your website and it pre-fills your organisation profile. Check it and fix anything wrong.
- Documents. Upload your latest accounts, strategy, impact report and two or three of your best recent bids.
- Past answers. Every answer you approve is stored, so the drafter and your connected AI can reuse your best wording next time.
- Your AI. Paste the MCP URL into Claude, ChatGPT or Cursor and ask it to summarise your profile. If the summary is right, it is ready to draft.
Who should pick which
- Pick Grant Assistant if you want a specialist model, an experienced international development team behind it, and a tailored contract.
- Pick GrantGunner if you want clear pricing and a workspace where your people and the AI tools they already use draft, review and hand over the same applications.
Learn more
See how GrantGunner works, pricing, what AI clients do with GrantGunner, and our roundup of the best AI grant writing tools in 2026.
Start free with GrantGunner and test it on a live question today.