AI assistants for writers and journalists
Which AI assistants can finish a job on the web, remember your project from one week to the next and answer an editor or a source by email?
38 assistants ranked
Thirty-eight assistants took at least two of the three tests that matter most to a writer's day: Online task, Memory and Email replies. Instinct leads at 9.7 over all three, then Hark and Dots at 9.0 over all three and Muse at 9.0 over two (it has no Memory score). Instinct and Hark score 10 on Memory; Instinct, Dots and Muse score 10 on Email replies.
Top picks
- 19.7InstinctTop pick
Online tasks 9Memory 10Email 10
- 29.0HarkBest free option
Online tasks 8Memory 10Email 9
- 39.0
Online tasks 9Memory 8Email 10
- 49.0
Online tasks 8Memory —Email 10
- 58.0TownFastest that's still good
Online tasks —Memory 7Email 9Time to finish 6m 30s
Top pick: highest average here. Best free option: highest among assistants you can use for free. Fastest that's still good: quickest median time to finish among those averaging 7 or more.
Full ranking
| # | Assistant | Online tasks | Memory | Average | |
|---|---|---|---|---|---|
| 1 | Instinct | 9 | 10 | 10 | 9.7 |
| 2 | Hark | 8 | 10 | 9 | 9.0 |
| 3 | Dots | 9 | 8 | 10 | 9.0 |
| 4 | Muse | 8 | — | 10 | 9.0 |
| 5 | Town | — | 7 | 9 | 8.0 |
| 6 | Ollie | 9 | — | 7 | 8.0 |
| 7 | Asmi | 7 | — | 9 | 8.0 |
| 8 | szn | 7 | 9 | 7 | 7.7 |
| 9 | Waldo | 6 | 8 | 9 | 7.7 |
| 10 | Bantr | 8 | 7 | — | 7.5 |
| 11 | Norton Family Assistant | — | 7 | 8 | 7.5 |
| 12 | Tomo | 7 | 7 | 8 | 7.3 |
| 13 | Neo | 7 | 8 | 6 | 7.0 |
| 14 | Poke | 4 | 8 | 9 | 7.0 |
| 15 | Caddy | 8 | 6 | 7 | 7.0 |
| 16 | Familiar | 7 | 6 | 8 | 7.0 |
| 17 | Pally | 7 | — | 7 | 7.0 |
| 18 | Orbits | 7 | 6 | — | 6.5 |
| 19 | Luke | 6 | 6 | 7 | 6.3 |
| 20 | Gemini Spark | 6 | 4 | 8 | 6.0 |
| 21 | Buoy | 7 | 5 | — | 6.0 |
| 22 | OpenInstinct | 10 | 4 | 3 | 5.7 |
| 23 | Today | 8 | 5 | 4 | 5.7 |
| 24 | Vellum | 4 | 7 | — | 5.5 |
| 25 | Nessi | — | 4 | 7 | 5.5 |
| 26 | Wajo / Fo | 2 | 7 | 7 | 5.3 |
| 27 | OpenClaw | 7 | 5 | 4 | 5.3 |
| 28 | Tressa | — | 5 | 5 | 5.0 |
| 29 | Persona | 5 | — | 5 | 5.0 |
| 30 | Ask Ranger | 4 | 6 | 4 | 4.7 |
| 31 | Del | 5 | 4 | — | 4.5 |
| 32 | Joshu | 5 | 3 | 5 | 4.3 |
| 33 | Manus | 4 | 4 | 4 | 4.0 |
| 34 | tinyNature | 4 | 3 | 5 | 4.0 |
| 35 | Lucas | 4 | 5 | 3 | 4.0 |
| 36 | The Shadow | — | 4 | 4 | 4.0 |
| 37 | Tab | 2 | 5 | — | 3.5 |
| 38 | AVA | 3 | 4 | 3 | 3.3 |
How this report works
Included are assistants with at least two of Online task, Memory and Email replies taken (catalog checked 2026-10-26), ranked by the average of the tests they took. Online task checks that it finishes a real job on a real website, Memory that it remembers, updates and lets you delete what it knows, Email replies that it answers on the right thread. We do not test research quality, source checking, quotes, drafting or editing prose, so this ranks how well an assistant handles the errands around the writing.
How to choose
- Read the Memory column first, since a long project needs the assistant to keep its background: Instinct and Hark score 10, szn 9, Neo, Waldo, Poke and Town 7 to 8. Muse, Ollie and Asmi took no Memory test, so their averages rest on two tests.
- For answering editors and sources, Email replies: Instinct, Dots and Muse score 10, Hark, Town, Poke, Waldo and Asmi 9. For work on the web, Online task: OpenInstinct scores 10, Instinct, Dots and Ollie 9.
- Check cost and access: Hark lists $20 a month, Town $15, Dots $100; Waldo and Familiar are on waitlists, and Instinct lists no price in our catalog.
Which AI assistant is best for writers?
Across Online task, Memory and Email replies, Instinct averages 9.7, Hark and Dots 9.0 and Muse 9.0 over two tests. We do not test writing quality.
Can an AI assistant remember my book or story project?
The Memory test checks that it remembers what you said, updates it when you change your mind and lets you see and delete it. Instinct and Hark score 10 on it.
Do you test research or fact-checking for journalists?
No. We test the errands around the work: finishing a job on a website, remembering context and replying to email. Check sources yourself.
What users say
Last week it told me 'NOT TO OPEN' an email from a friend inviting me to a backyard BBQ—it turned out to be a phishing scam.
Open sourcePeople have reported that when they tried to disconnect the agent from their email, they found out it was retaining a copy of their inboxes anyway.
Open sourceThe agent, which Meta launched on September 8, booked a Friday date night and made a packing schedule for my upcoming move. It also emailed a friend about a trip next month and made a planning document with places we want to visit.
Open sourceFirst, I asked it to look through my email and send me a daily digest of "important" emails (my inbox gets quickly clogged with PR pitches and newsletters). It did this admirably well.
Open sourceFrequently asked questions
Which assistant is best for this?
Instinct leads with an average of 9.7, ahead of Hark and Dots.
Is there a good free option?
Hark is the highest-ranked assistant with a free plan, averaging 9.0.
How did you test these assistants?
Every assistant got the same published tests, word for word, scored from three checkpoints and the number of corrections it needed. Each test is on the Dimensions page.