Opportunities › Open-source durable compute for AI agents › Proposed app
Hearth Actors
Self-hosted stateful actors with built-in storage for multiplayer AI agents
- Score
- 37
- confidence 53
- Evidence
- 9
- items on the page
- Rivals
- 4
- named in sources
- First version
- 60 d
- estimate
- Type
- Desktop app
- PWA: runs in the browser, installs on Mac and Windows
- Market since
- 2020
- Established market · AI estimate
ScreensConcept
Desktop app: Infrastructure engineers inspect actor state, set compute limits, and manage fleet configuration in a sustained monitoring dashboard. Sample screens on a computer, in the proposed brand. Their numbers are sample data.
Fleet
All nodes healthy1,284
About this appAI proposal
Hearth Actors gives each agent, room or session its own isolated actor with its own database, memory and compute limits. You run it on your own servers or cloud account, so state lives where you choose and your agents are not tied to one provider's storage and edge products.
It is for AI teams and infrastructure engineers building multiplayer agent systems. Instead of pushing every update through one central database and a websocket server, each actor owns its state and handles its own concurrent calls, and actors spread across nodes as load grows.
The core is open source. A small PWA dashboard shows running actors, their state size and their connections, so you can inspect and debug an agent fleet from a phone or a laptop. Paid plans cover hosted control plane and support.
- Tagline
- Give every agent a home you own.
- Voice
- PlainPracticalCalm
- Why this name
- A hearth is a fixed place that keeps a fire going. Each actor keeps its state alive in a place you control.
In the first version
Members say which features they would use. It tells PWAS what people want; it changes no score.
Isolated actors with their own database
Every actor gets private embedded storage. Reads and writes are serialized per actor, so concurrent calls do not corrupt state.
No member yet
Horizontal placement across nodes
Add nodes and actors are placed and moved between them. No single database or websocket server carries all traffic.
No member yet
Configurable compute per actor
Set memory, CPU and idle timeout per actor class, so a heavy planner agent and a light chat room can differ.
No member yet
Live connections built in
Clients and other agents connect to an actor by ID over websockets, so multiplayer sessions share one source of truth.
No member yet
Brand identityAI proposal
Check the name against trademark registers before using it.
Palette
#C2410CEmber · primary#FBF6EFWarm Paper · surface#1F1A17Soot · text#3F6B4FMoss · accent, healthy status#E7D8C6Clay · border, secondary surface
Typography
Aa
Hearth Actors
ABCDEFGHIJKLMNOPQRSTUVWXYZ 0123456789
Give every agent a home you own.
PricingAI proposal
The material gives no price for durable compute. Liveblocks, a related multiplayer product, lists paid plans from $25 to $500 a year plus a free plan; these tiers are a starting point to test, not a match.
Open source
Free
- Self-hosted actor runtime
- Embedded per-actor database
- Community support
Team
$25 a month
- Hosted dashboard and control plane
- Snapshots and alerts
- Email support
Scale
$500 a month
- Multi-node fleet management
- Priority support
- Deployment review
Would you pay Team · $25 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. It reaches customers through open-source repository visibility, documentation, example agent templates and developer content. Runs inside customer infrastructure, so end users do not see it.
- Do they become customers?
- Developers who find the repo can become paying users of hosted or support tiers. Test whether self-hosting teams convert to the paid control plane or support.
- First customers
- Small AI startups of 3 to 15 engineers building multiplayer or multi-agent products on Cloudflare Durable Objects who want to move off it.
- Where to reach them
- Hacker News: Post a Show HN with a migration guide and a benchmark of a multiplayer agent demo against a central database setup.
- Open-source repository: Publish the runtime with agent templates and a quickstart so developers can deploy in under an hour.
- AI engineer communities: Join Discords and forums for agent builders, answer state and concurrency questions and offer migration help.
- First milestone
- Five teams run actors in production and two pay for the hosted dashboard or support within 90 days of release.
Why people already pay for this
- VerifiedHacker News post received 41 points and 25 comments
- Public claimDurable Objects from Cloudflare create vendor lock-in and bundle unnecessary services
- Public claimDurable Objects were not originally designed for AI agent workloads
- Public claimDurable Actors are needed for deploying multiplayer agents with horizontal scalability
- 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 · Actor runtimeSelf-hostable server runtime with embedded SQLiteEach actor needs isolated state and serialized execution, which Postgres and Vercel do not provide.
- This app · Node placementLightweight cluster coordinatorRoutes actor IDs to nodes and moves actors when nodes join or leave.
RisksAI inference
- TechnicalReliable actor placement, failover and state migration across nodes is hard; bugs can lose or duplicate state, and trust in a state layer is slow to earn.
- CommercialOpen-source infrastructure is hard to monetize. Teams may self-host forever and never buy the hosted or support tiers.
- PlatformCloudflare's incumbent product is widely used and already has multiplayer frameworks built on it. Switching costs may outweigh lock-in worries.
- DataAgent state may hold sensitive user data. Customers will expect encryption, access controls and clear backup handling from the start.
- CommercialDemand evidence is a single Hacker News post (41 points, 25 comments) and public claims; real willingness to pay is unverified.
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.