Four AI wrappers: $70–$3,150 MRR, and the setup tax (2026)

[data] Four small AI wrappers cleared this week's test: public recurring revenue, a 2026 founding or launch signal, and a workflow narrow enough for an indie developer to understand. The revenue snapshots come from TrustMRR as displayed on August 31, 2026. 1 2 3 4

[editor's view] TrustMRR's founding month is a timing proxy, not an exact product launch date. Model access is cheap. Setup, context, and repeat use are the work.

TL;DR

[data] Coral is the largest business in this set at $3,150 MRR and 44 active subscriptions, with Stripe-verified revenue on TrustMRR; its product packages OpenClaw deployment, hosting, integrations, and templates. 1 [data] StartClaw follows at $1,619 MRR and 21 active subscriptions, while its founder publicly reported 700 signups and 50 paying users during the February launch. 2 5 [data] Stripe Report shows $70 MRR and seven active subscriptions for a narrow reporting job. 3 [data] LiveCrew AI is the smallest signal at $58 MRR and two active subscriptions, selling four persistent AI executive roles to solo founders. 4

[editor's view] These products make four bets on wrapper economics. Coral and StartClaw charge for getting an agent running. LiveCrew charges for structured disagreement. Stripe Report charges for a forward-looking operating view. The safest clone test is to manually deliver the outcome to five target users before paying for a large automation layer.

Three takeaways

[data] Distribution appears before moat. Coral's founder reported 10-plus workshops in one month, and StartClaw's founder used X plus a Discord help channel during launch. 6 7
[editor's view] Setup is the product when the underlying agent is easy to copy. Hosting, credentials, templates, integrations, and onboarding can create more customer value than a thin prompt layer.
[editor's view] The second paid use matters more than the first signup. A free trial can create a spike; a repeated job proves that the workflow earns a place in the user's week.

Replication scoreboard

[editor's view] Scores run from one to five stars. More stars mean more work or exposure; legal risk measures exposure from data, claims, and permissions.

Product	Narrow workflow	MRR / active subscriptions / snapshot	Best public acquisition evidence	Technical lift	Information edge	Capital needed	Legal risk	Clone read
Coral	Deploy and operate OpenClaw across business tools	[data] $3,150 / 44 / Aug. 31, 2026 1	[data] Founder-led workshops plus X promotion; revenue attribution is undisclosed 6	★★★★☆	★★★☆☆	★★★☆☆	★★★★☆	[editor's view] Valuable setup layer; permissions and reliability are the wall
LiveCrew AI	Pressure-test a solo founder's product or launch decision	[data] $58 / 2 / Aug. 31, 2026 4	[data] Founder-led X and Product Hunt launch activity 8	★★★☆☆	★★★☆☆	★★☆☆☆	★★☆☆☆	[editor's view] Easy to prototype; hard to make disagreement useful
Stripe Report	Forecast cash and revenue risk from subscription data	[data] $70 / 7 / Aug. 31, 2026 3	[data] Primary channel not publicly disclosed	★★★☆☆	★★★★☆	★★☆☆☆	★★★★☆	[editor's view] Small build with a trust-heavy data boundary
StartClaw	Launch a hosted OpenClaw agent without server or API-key setup	[data] $1,619 / 21 / Aug. 31, 2026 2	[data] Founder-led X, Discord support, and an early referral/community loop 7	★★★★☆	★★★☆☆	★★★☆☆	★★★★☆	[editor's view] The onboarding win is clear; platform dependence is clear too

Coral: OpenClaw without the weekend setup

[editor's view] Positioning: Coral is a managed deployment layer for OpenClaw agents that connects an always-on agent to the tools a small business already uses.

[data] Coral's TrustMRR page lists $3,150 MRR, 44 active subscriptions, and $41,291 in all-time revenue. The page says revenue was verified through Stripe and was last updated August 31, 2026. TrustMRR lists Coral as founded in February 2026; that month is the available timing proxy. 1

[data] The page lists Standard at $50 per month and Ultra at $200 per month. Coral's official site says deployment takes minutes, provides isolated environments and templates, connects common work tools, and routes tasks across OpenAI and Anthropic models. 1 9

[data] Job-to-be-done: a founder wants an AI employee running inbox triage, research, scheduling, drafting, or monitoring without buying a server or maintaining an agent runtime. Coral's official workflow copy supports those use cases. 9

[data] Coral's founder said the company hosted 10-plus workshops in one month, helped thousands deploy OpenClaw, and saw users spending more than $300 per day on the platform. Those are founder claims; the post does not attribute paid revenue to workshops or X. 6

[editor's view] The cloneable part is the five-minute start. The hard part is the permission graph: every integration adds OAuth edge cases, data isolation, recovery, and support. Coral's three named co-founders fit the team ceiling, while TrustMRR has no formal team-size field. 6

[editor's view] Replication score: Technical lift ★★★★☆ because a reliable multi-integration runtime is more than an API call. Information edge ★★★☆☆ because workflow templates and support knowledge accumulate, while OpenClaw remains available to competitors. Capital needed ★★★☆☆ because hosting and model usage grow with active agents. Legal risk ★★★★☆ because the product touches email, calendars, customer records, and code repositories.

If you wanted to copy this

[editor's view] First step: sell one managed agent for one tool pair, such as Gmail plus Slack, and manually handle credential setup for five businesses.
[editor's view] First likely failure mode: the agent takes an unsafe action or silently fails after an integration token expires, turning the promised time savings into a support queue.

LiveCrew AI: the product is disagreement

[editor's view] Positioning: LiveCrew AI sells four persistent AI executives—a CTO, CMO, Head of Product, and Head of Operations—as a structured decision room for a solo founder.

[data] TrustMRR lists $58 MRR, two active subscriptions, and $141 in all-time revenue, verified through Paddle and updated August 31, 2026. The page lists April 2026 as the founding month, a timing proxy, a $29 monthly price, a one-person team, and bootstrapped funding. 4

[data] LiveCrew's official site says a user briefs the four roles, runs meetings or one-on-one threads, and receives recaps and versioned documents. The agents retain meeting context and can push back on earlier decisions. The site lists a seven-day trial followed by $29 per month. 10

[data] Job-to-be-done: a solo founder wants to pressure-test a product, pricing, or launch decision and leave with a decision record plus working documents. Role-specific prompts, personas, threads, and memory are the disclosed mechanics. 10

[data] The founder announced LiveCrew as live on X on August 22, 2026, describing the four executives and $29 monthly beta pricing. Public launch activity also points to Product Hunt. Public material gives no channel-attributed conversion data. 8

[editor's view] LiveCrew is approachable because a first version can be four role prompts, shared memory, and a document exporter. The product risk is behavioral: agreeable bots can create the feeling of a leadership team without useful friction. Two active subscriptions leave retention unresolved.

[editor's view] Replication score: Technical lift ★★★☆☆. Information edge ★★★☆☆ because a good founder-decision corpus could improve the prompts, but role labels are easy to copy. Capital needed ★★☆☆☆ for a text-first prototype. Legal risk ★★☆☆☆ because the product mainly handles founder-entered business plans, although confidentiality still matters.

If you wanted to copy this

[editor's view] First step: run ten paid decision reviews by hand, with one human facilitator switching between four written roles, and save the documents that customers actually use.
[editor's view] First likely failure mode: users enjoy the meeting transcript but never return for a second decision because the disagreement produces theater rather than a better choice.

Stripe Report: a forecast layer on top of Stripe

[editor's view] Positioning: Stripe Report turns subscription billing events into forecasts, cancellation alerts, MRR trends, and recurring reports.

[data] TrustMRR lists $70 MRR, seven active subscriptions, and $170 in all-time revenue. Stripe verified the revenue through an API key, and the page was updated August 31, 2026. TrustMRR lists April 2026 as the founding month, a timing proxy. The page lists launch-special pricing of $9.99 per month after a seven-day trial, then $19.99 and $49 plans. 3

[data] The official site says Stripe Report covers expected revenue, pending cancellations, month-over-month trends, and email reports. TrustMRR identifies Stripe and Anthropic in the backend stack. The site's dashboard figures are examples, so this article uses the TrustMRR snapshot for traction. 3 11

[data] Job-to-be-done: a subscription business owner wants to know expected cash, pending cancellations, and revenue at risk without exporting Stripe data into a spreadsheet and maintaining the formulas. The official product copy names those questions and outputs. 11

[data] The primary acquisition channel is not publicly disclosed. The public listing supplies a low domain-rating signal and a product page, but it does not establish which source produces paid customers. 3

[editor's view] Stripe Report has the cleanest wedge here. The clone is a small dashboard; the business is trust in the definitions. Forecasts need rules for renewals, pauses, failed payments, annual plans, and cancellations. Anthropic can summarize an account, but a model should not decide what counts as recognized revenue. The public page says the founder is looking for a co-founder, supporting a solo-founder reading while leaving the formal team boundary unresolved. 3

[editor's view] Replication score: Technical lift ★★★☆☆. Information edge ★★★★☆ because correct billing edge cases and trustworthy benchmarks matter. Capital needed ★★☆☆☆ for a read-only Stripe integration. Legal risk ★★★★☆ because financial reporting errors and access to customer billing data can damage trust.

If you wanted to copy this

[editor's view] First step: build a read-only report for one subscription business and reconcile every line against its Stripe dashboard for a full billing cycle.
[editor's view] First likely failure mode: the forecast looks precise while mishandling an annual plan, a paused subscription, or a failed renewal.

StartClaw: one-click OpenClaw hosting

[editor's view] Positioning: StartClaw makes a hosted OpenClaw agent available without the server setup or separate AI API key.

[data] TrustMRR lists $1,619 MRR, 21 active subscriptions, and $31,218 in all-time revenue, with Stripe-verified revenue and an August 31, 2026 update. The page lists January 2026 as the founding month, a timing proxy, and names Idan Mann as founder. The page lists Basic at about $8 per month, Pro at about $24, Max at about $166, and a custom plan. 2

[data] Idan Mann's February 6 launch post described hosted OpenClaw with no API key, offered a 48-hour trial, and pointed users to Discord. The next day, Mann reported 700 signups and 50 paying users, while noting that free trials with AI credits may have contributed. These founder disclosures are separate from the August TrustMRR snapshot. 5 7

[data] Job-to-be-done: a user wants an OpenClaw agent live in about a minute from a phone or browser, without configuring a server, managing deployment, or entering an AI provider key. The launch post supports the hosted and no-key promise. 7

[data] The best evidenced acquisition path is founder-led X combined with Discord support and community sharing. Public material shows the distribution surfaces and early signup and paying-user claims; it does not attribute the current MRR to one channel. 5

[editor's view] StartClaw shows why an onboarding shortcut can monetize before an agent workflow is differentiated. The first value arrives at deployment. The long-term question is whether users keep paying after the novelty fades. The disclosed web and cloud stack makes the shell reproducible; reliable agent operation remains harder. 2

[editor's view] Replication score: Technical lift ★★★★☆. Information edge ★★★☆☆ because onboarding and support knowledge compound, while the core deployment promise is legible. Capital needed ★★★☆☆ because hosted agents consume infrastructure and model budget. Legal risk ★★★★☆ because the service runs user-controlled agents and may touch connected accounts.

If you wanted to copy this

[editor's view] First step: pick one agent template, charge for a managed deployment, and watch five users complete the setup from a phone.
[editor's view] First likely failure mode: free credits attract experiments that never become recurring workloads, leaving hosting costs behind each signup.

The pattern: sell the setup, then earn the second use

[editor's view] Coral and StartClaw sell deployment. LiveCrew sells a decision ritual. Stripe Report sells an operating report. Each puts a layer between a general model or agent framework and a task with a buyer.

[editor's view] Test the second use, not the first moment of curiosity. Give five target users a manually assisted result, charge for it, and ask each buyer to buy the result again. That reveals whether the wrapper owns a recurring job or packages a demo.

Sources

Source	What it supports
[data] Coral on TrustMRR	MRR, subscriptions, founding proxy, price, verification
[data] Coral official site	Deployment, integrations, models, use cases
[data] Coral founder workshop post	Founders, workshops, and reported spend
[data] LiveCrew AI on TrustMRR	MRR, subscriptions, founding proxy, team, price
[data] LiveCrew AI official site	Roles, meetings, memory, documents, and trial price
[data] LiveCrew AI launch post	August 22 launch and beta price
[data] Stripe Report on TrustMRR	MRR, subscriptions, founding proxy, price, stack
[data] Stripe Report official site	Forecast, cancellation, trend, and report workflow
[data] StartClaw on TrustMRR	MRR, subscriptions, founding proxy, price, founder, stack
[data] StartClaw launch post	Hosted/no-key launch, 48-hour trial, and Discord
[data] StartClaw launch and early users	Reported 700 signups and 50 paying users

What to do tomorrow morning

[editor's view] Pick one narrow workflow from this issue, find five people who already perform it, and offer a manually assisted version for money. Ask each buyer to pay for the same result a second time before you build hosting, integrations, or a polished agent layer.