Opportunities › AI-powered code completion for Neovim › Proposed app
Keelstroke
Fast, context-aware AI completion that lives inside Neovim
- Score
- 37
- confidence 47
- Evidence
- 8
- items on the page
- Rivals
- 4
- named in sources
- First version
- 45 d
- estimate
- Type
- Desktop app
- PWA: runs in the browser, installs on Mac and Windows
- Market since
- 2021
- Emerging market · AI estimate
ScreensConcept
Desktop app: Developers work in their editor for extended periods writing and editing code, needing inline suggestions as they type. Sample screens on a computer, in the proposed brand. Their numbers are sample data.
Today
Acceptance rate68%
About this appAI proposal
Keelstroke is a Neovim plugin that suggests your next edit as you type and lets you accept it with one key. It aims for the tab-completion feel that people now leave Neovim to get, without asking you to change editors, keymaps or terminal.
A small web app, installable as a PWA, is the control room: sign in, pick models, set per-project rules, watch usage and manage billing. Your editor stays plain Lua and your config stays in your dotfiles.
It is built for modal editing first. Suggestions respect normal, insert and visual modes, sit beside your existing completion menu, and can be switched off per filetype or repo.
- Tagline
- Tab-completion quality, without leaving Neovim.
- Voice
- DirectCalmTechnical
- Why this name
- A keel keeps a boat steady on its course; a keystroke is the unit of editing. Together they suggest steady, fast typing in a terminal.
In the first version
Members say which features they would use. It tells PWAS what people want; it changes no score.
Next-edit suggestions
Predicts the next change as ghost text, including multi-line edits near the cursor, and accepts it with a single mapped key.
No member yet
Mode-aware by design
Works in insert, normal and visual modes. An accepted suggestion is one undo step and repeats with dot.
No member yet
Works beside your completion menu
Sits next to LSP and snippet completion instead of replacing them, with priority rules you can change.
No member yet
Per-project rules
Set model, context limits and ignore patterns for each repo in the web app or in a local config file.
No member yet
Brand identityAI proposal
Check the name against trademark registers before using it.
Palette
#14181ADeep Slate · surface#E8EDE6Paper Mist · text#7FD1A8Terminal Mint · primary#F2B84BAmber Key · accent#6FA8DCHarbor Blue · data#2A3236Moss Grey · panel
Typography
Aa
Keelstroke
PricingAI proposal
Based on pricing pages in the material: Cursor $20 to $40 a month, Copilot $4 to $21, Supermaven $10, all with a free plan. Tiers sit inside that range and are a starting point to test.
Free
Free
- Limited completions each month
- Fast default model
- Plugin and web app
Pro
$15 a month
- Much higher completion limits
- Best suggestion model
- Per-project rules
- Request log and stats
Power
$30 a month
- Highest limits
- Choice of several models
- Longer context
- Priority speed
Would you pay Pro · $15 a month?
Your answer tells PWAS which prices people accept. It changes no score.
No answers yet.
What it could be worthEstimate
What a software business making this much sells for today, from real deals. Pick a monthly revenue.
If it makes
Worth about
$194,000
Sale range $132,000 to $263,000
$120,000 a year in revenue
From 87 software businesses listed and sold on Empire Flippers in the last year, revenue vetted by the broker: they sell for about 19× their monthly revenue, most between 13× and 26×.
Measured on profit, the middle deal is 32× monthly net profit: about 2.7 years of profit.
An estimate, not an offer or financial advice. A real price depends on profit, growth, churn, how much of the work is the owner's, and the buyer.
How it would growAI proposal
How the product would reach new customers through its own use, and how the first ones would be found. A plan to test, not a result.
- Built-in loop
- No loop in the product. A plugin installed from a repo spreads by word of mouth, dotfile sharing and README mentions, but the product does not itself send anything to people who do not use it.
- Do they become customers?
- Reach comes from content, community posts and plugin lists. Test whether free-tier users convert to Pro after two weeks of daily use, and whether dotfile mentions bring installs.
- First customers
- Professional developers who use Neovim daily and already pay for or have tried an AI assistant, working in one or two languages such as Go, Rust or TypeScript.
- Where to reach them
- Hacker News: Post a Show HN with a short demo and an honest latency comparison, and answer comments about Neovim AI tools.
- Neovim communities: Share in the Neovim subreddit and Matrix or Discord channels, and ask for early testers with feedback on a form.
- Dev blog and videos: Publish a setup guide and a screen recording of a real editing session using the plugin.
- First milestone
- Fifty weekly active free users and ten paying Pro subscribers within eight weeks of launch, with acceptance rate above 50%.
Why people already pay for this
- Public claimDevelopers would pay for a comparable Neovim AI code completion solution
- Public claimExisting AI coding assistants for Neovim are significantly inferior to proprietary editor alternatives
- AI inferenceThe lack of a comparable AI tool for Neovim is causing developers to switch editors despite preferring Neovim
- Public claimThere is a significant community of Neovim users facing this tradeoff
- Verified
Tech stack
Every proposal builds on the same PWA base, installable from the browser on any phone or computer. Dashed boxes are what this app adds.
- AppNext.js and TypeScript
- PWAManifest and service worker
- UITailwind CSS
- Data and authSupabase Postgres
- PaymentsStripe
- EmailResend
- HostingVercel
- This app · Editor pluginLua plugin for NeovimThe product lives in the editor: ghost text, keymaps, mode handling and request batching.
- This app · Inference gatewayStreaming API over hosted code modelsLow latency suggestions need streaming, caching and model routing, beyond ordinary serverless routes.
RisksAI inference
- AIMatching the quality of leading editors depends on a model tuned for next-edit prediction. Off-the-shelf models may feel slower or less accurate, and users judge it within minutes.
- TechnicalGhost text, multi-line edits and undo behaviour in Neovim are fiddly across versions and plugin setups; latency over the network may not feel instant.
- CommercialEvidence that developers would pay is a public claim from Hacker News, not verified. Free tiers from Copilot and Supermaven make paying for a Neovim-only tool a hard sell.
- DataSending code to a hosted service worries many terminal users. It needs clear retention terms, exclusion rules and possibly a bring-your-own-key option.
- PlatformLarge vendors could improve their own Neovim support, and model providers can change prices or terms, squeezing margins.
Everything marked AI proposal or AI inference was written by a model from the evidence on the opportunity page. The proposed name must be checked against trademark registers before it is used.