Opportunities › Local-first AI agent for terminal operations › Proposed app
Tillhand
A terminal agent that runs on your machine and asks before every command
- Score
- 45
- confidence 41
- Evidence
- 5
- items on the page
- Rivals
- 3
- named in sources
- First version
- 45 d
- estimate
- Type
- Desktop app
- PWA: runs in the browser, installs on Mac and Windows
- Market since
- 2023
- Emerging market · AI estimate
ScreensConcept
Desktop app: Developers work in a terminal approving AI-proposed commands on their machine, with phone approvals as a secondary feature. Sample screens on a computer, in the proposed brand. Their numbers are sample data.
Approval needed
Waiting for you1 pending
About this appAI proposal
Tillhand is an AI agent that works in your terminal on your own computer. It proposes commands, you approve or decline each one, and it runs them locally. Your commands and output stay on your machine unless you choose otherwise.
Built for developers and system administrators who want autonomous help without sending system operations to a third party. Every action is written to an audit log in a standard format, so you and your team can review what happened and why.
The web companion is a PWA for reviewing approvals, browsing logs and managing team policies. The agent itself runs locally on macOS, Linux and Windows.
- Tagline
- Autonomous help. Your machine. Your call.
- Voice
- CalmDirectCareful
- Why this name
- Swedish-sounding 'till hand' means 'at hand': the agent is close, on your own machine, and waiting for your word.
In the first version
Members say which features they would use. It tells PWAS what people want; it changes no score.
Approve every action
Each command shows what it will do and where. Approve once, approve for this session, or decline. Nothing runs without a rule or a click.
No member yet
Runs on your machine
The agent executes locally. Command text and output are not sent to a Tillhand server.
No member yet
Standard audit logs
Every proposal, approval and result is recorded as JSON lines you can export to the tools you already use.
No member yet
Policy rules
Allow safe commands such as read-only listings, require approval for writes, and block patterns like recursive deletes.
No member yet
Brand identityAI proposal
Check the name against trademark registers before using it.
Palette
#14352FDeep Pine · primary#F6F3ECPaper · surface#1B1F1EInk · text#E0902BAmber · accent#5E8C6AMoss · data
Typography
Aa
Tillhand
ABCDEFGHIJKLMNOPQRSTUVWXYZ 0123456789
Autonomous help. Your machine. Your call.
PricingAI proposal
The material gives no competitor prices. These tiers are a starting point to test, following the freemium and per-seat model in the opportunity.
Local
Free
- Local agent, open-source core
- Per-command approvals
- Local audit log
- Basic policy rules
Pro
$12 a month
- Phone approvals via PWA
- Log export and retention
- Unlimited policy rules
- Multiple devices
Team
$25 a month
- Per seat price
- Shared policies
- Approver roles
- Compliance log formats
Would you pay Pro · $12 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
- Partners and integrations. The agent plugs into editors, shells and existing coding assistants, and exports logs to tools teams already use. Its open-source core spreads through repositories and developer communities.
- Do they become customers?
- Open-source users can become paying customers when they need team policy, phone approvals or compliance logs. Test whether individual users convert, or only teams.
- First customers
- Small engineering teams of 5 to 30 people at privacy-sensitive software companies, led by a platform or security-minded engineering lead.
- Where to reach them
- GitHub open-source launch: Release the local core with a clear README, a demo and a good-first-issue list.
- Developer forums: Post a technical write-up of the approval and audit design in relevant communities.
- First milestone
- Five teams paying for shared policies and audit export within 90 days of the open-source release.
Why people already pay for this
- VerifiedGitHub repository gained 424 stars in 291 days
- Public claimExisting commercial alternatives operate on cloud infrastructure, raising privacy and control concerns for users who need autonomous system access.
- Public claimUsers need explicit approval mechanisms for autonomous terminal operations to maintain safety and auditability.
- AI inferenceThe market for AI-powered development agents is active with multiple competing products.
- Public claim
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 · Local agentCross-platform desktop daemon and CLICommands must execute on the user's machine on macOS, Linux and Windows.
- This app · Model runtimeLocal model runner plus optional own API keyKeeps inference on the machine by default; users may bring their own provider.
RisksAI inference
- TechnicalSafely sandboxing and classifying commands across three operating systems is hard; a missed dangerous command could damage a user's machine.
- AISmall local models may plan worse than cloud models, so users may find the agent less capable than the cloud tools.
- CommercialFree open-source alternatives such as Goose exist, so paid features must be clearly worth paying for.
- PlatformApprovals relayed to a phone need a sync service, which must carry only metadata to keep the privacy promise.
- CommercialCloud vendors are adding autonomous agents quickly and could add local or approval modes.
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.