# SuperPenguin (Carrot Labs) Aiolum story: https://aiolum.com/story/superpenguin/ · added 2026-11-12 · updated 2026-11-12 **In short.** SuperPenguin (Carrot Labs) (AI cost management: attribute AI spend by customer, feature, team and pull request, USA): Self-serve paid plans listed, November 2026. Evidence grade E: no revenue figure was checked. Team: 2 founders (Y Combinator company page, Winter 2026 batch). Launched April 2026; still running. AI cost management: attribute AI spend by customer, feature, team and pull request · USA · AI-native · team of 2 to 10 · still running · group: Developer tools and coding agents - Outcome: Self-serve paid plans listed, November 2026 (evidence grade of the headline number: E; no revenue figure) - Launch: April 2026; months from launch to the outcome: 7 - Team: 2 founders (Y Combinator company page, Winter 2026 batch) - Outside funding: Y Combinator, Winter 2026 batch; amount not published - Main factor: Product - Meters, 1 to 4: evidence strength 1, copyable pattern 2, capital needed 2 - Tags: AI-native, Small team, Funded, Running, B2B, Developer tools Links: - Visit superpenguin.ai: https://superpenguin.ai/ **The product.** Software that shows what a company's AI bill was spent on. It reads the bills of AI providers and tags each call with a customer, a feature or a team. It also prices each pull request that an AI coding tool helped write. The company says it records cost data only and does not sit in the request path. **Why it worked.** The company says provider dashboards show only their own bill, by project, model and API key. They cannot group spend by a customer's own customer, feature or pull request. SuperPenguin sells that missing view. Its pricing page lists a free plan up to US$2,000 of managed spend and paid plans above it. Evidence grades: A = disclosed by a public company, a filing or a payment processor, or verified by one; B = two or more independent reports agree, or a journalist confirmed it; C = a founder claim that no one else verified; D = reported, not verified: one founder claim or a third-party estimate with no second source; E = no revenue figure found, so the headline is a non-money fact and the other facts carry their own grades; calc = computed on this page from graded facts. ## Summary SuperPenguin is an AI spend attribution platform: it splits an AI bill into the customers, features, teams and pull requests that caused it. A pull request is a proposed code change that a team reviews before it is merged. The product has Python and TypeScript SDKs (code libraries that a team adds to its own program) to tag each AI call. It also has a Mac app and a connection to GitHub for the coding view. The company is Carrot Labs AI, Inc., and two founders are named on its Y Combinator page: Christopher Acker and Yuta Baba. The earliest dated copy of the Carrot Labs site is from 5 February 2026 and shows two other tools. On 6 March 2026 the company posted a different product on Y Combinator, a service that tunes models to a customer's tasks. The first dated SuperPenguin page, the privacy policy, is effective 12 April 2026. No revenue figure was found (grade E). What is checked is that paid plans are on sale. The pricing page lists Growth at US$30 and Pro at US$200 with sign-up links. The terms describe fees at checkout, but no independent source says that a customer has paid. ## Why did SuperPenguin (Carrot Labs) work? ### Provider dashboards show their own bill, not the customer's customers The company says a provider dashboard shows its bill by project, model and API key. In the company's words it cannot show other providers, a team's own customers or features, or what a pull request cost. SuperPenguin joins the bills of several providers and adds those tags. This is the company's description of the gap, not an independent test. ### A free plan up to US$2,000 lets a team start without paying The pricing page lists a Free plan at US$0 for up to US$2,000 of managed spend. The page says that no card is needed and that Free is a plan, not a countdown. Growth at US$30 and Pro at US$200 add members, alerts and more spend. ### The company now presents SuperPenguin as its current product The Carrot Labs home page says that SuperPenguin is its current answer to where AI money goes. The February 2026 copy of the same site offered two other tools. The March 2026 Y Combinator post offered model tuning. The sources do not say why the focus changed. ## How it was built These details come from the company's own pages, its documentation index and its Y Combinator page. They are grade C unless the note says otherwise. - Python and TypeScript SDKs tag each AI call by customer, feature or team. The company says that the SDK sends usage data in the background while requests go straight to the AI provider. - A connection to a provider's admin key pulls the bill into one total. The Y Combinator page says that the product tracks 14 providers. - A free Mac app reads local usage files of Cursor, Claude Code and Codex. GitHub is needed to match that spend to merged pull requests. - By default only cost data is stored: token counts, model, latency and tags. The company says that prompt text is stored only if an owner turns capture on. - Stripe processes card payments. The privacy policy also says that an email address is looked up at Apollo.io, a business contact data provider, when an account is created. Sources: [1, 3, 9, 10, 13] ## Who tried this before SuperPenguin (Carrot Labs), and why did they fail? No record in our database of failed startups is about AI spend. These are three nearby cases: B2B analytics products that promised insight and then lost to larger vendors. The values are from the database on 12 November 2026 and are not independently verified. The failed startups come from our database of dead startups, retold here in our own words; Its entries are AI-assisted summaries of public sources, so they carry no evidence grade. ### Lariat Data (2021 to 2024; funding US$5.0M burned; cause: Competition) - Promise: A platform that gathered media data and turned it into insight for engineering, product and design teams. - Why it died: The page says strategic misalignment and intense competition, as Tableau and Looker added AI. - Record: ### Watchsend (2013 to 2016; funding US$120K burned; cause: Competition) - Promise: Analytics that showed developers how users behave inside a mobile app. - Why it died: The page says it struggled to stand apart from Mixpanel and Google Analytics. - Record: ### Teamometer (2015 to 2018; funding US$1.5M burned; cause: No market need) - Promise: Performance metrics that help a team improve how it works. - Why it died: The page says there was no proven market need, and organizations resisted change. - Record: ## What changed The three records sold insight about a team's work to buyers who already had cheaper or larger tools. SuperPenguin sells the same kind of insight about one cost: AI usage across providers. The company's own blog says that AI spend is now a real line item for teams. The risk is the same as in those records. The company's own blog roundup says that three kinds of tools touch LLM cost: gateways, observability tools and cost management tools. The sources here do not show how many customers pay. ## Can one person copy SuperPenguin (Carrot Labs) now? **Verdict: The structure can be copied; the proof of demand cannot be seen yet.** The pricing model is the part a reader can study. It has a free plan and two self-serve plans, priced by the AI spend under management, not by seats. No source shows that this earns money. A reader should treat the story as a design to study, not as evidence of a result. 1. Pick one cost that the providers do not split for the buyer, such as cost by customer. 2. Offer a free plan with a clear spend limit and no card. 3. Price the paid plans by the spend the product manages, as the company's page does. 4. Show a named customer story with a measured saving, and say who stated it. **What is still open.** Other costs that bills do not split by cause: search APIs, speech models, and the GPU time of agents. ## How much revenue does SuperPenguin (Carrot Labs) make? The numbers, graded A two-founder company with listed paid plans in November 2026 and no revenue figure. Every row comes from the company's own pages or its Y Combinator page | Fact | Value | Grade | Note | Sources | |---|---|---|---|---| | What it sells | AI spend attribution and coding cost per pull request | C | The company's FAQ: it turns spend across AI providers and coding tools into cost that a team can attribute and control. | [1] | | Paid plans | Growth US$30 and Pro US$200 per month; Enterprise custom; Free US$0 | C | Pricing page. The Free plan covers up to US$2,000 of managed spend and the page says it is not a trial. | [2] | | Self-serve path | Sign-up links for Growth and Pro, fees at checkout, Stripe named | C | The pricing buttons link to /signup?plan=growth and /signup?plan=pro. The terms say fees are stated at checkout. The privacy policy says Stripe processes card details. | [2, 8, 9] | | Team | 2 founders | C | Y Combinator company page: team size 2, with Christopher Acker as Founder/CEO and Yuta Baba as Founder. | [3] | | Company founded | 2026; exact date not found | C | Y Combinator lists the founding year as 2026. The earliest dated copy of carrotlabs.ai is from 5 February 2026 and names Carrot Labs AI, Inc. | [3, 5] | | Earlier tools of the same company | 3GPP Scout and an Earnings Excel plugin, seen 5 February 2026 | C | The archived carrotlabs.ai home page of 5 February 2026 lists these two products. | [5] | | Earlier model-tuning product | Posted on Y Combinator on 6 March 2026 | C | The post says Carrot Labs builds models tuned to a customer's tasks that keep improving. The post's creation time in the page data is 2026-03-06. | [4] | | First dated SuperPenguin page | Privacy policy effective 12 April 2026 | C | The policy says Carrot Labs AI, Inc. operates SuperPenguin. It was last updated on 22 October 2026. | [9] | | Funding | Y Combinator, Winter 2026 batch; amount not published | C | The home page shows a Backed by Y Combinator logo, and the Y Combinator page lists the batch. | [1, 3] | | Named customer stories | 2: Channel3 and Freebuff | C | Both are case studies on the company's own site. The Channel3 page says SuperPenguin found about a fifth of its AI spend in looping calls. | [10, 11] | | Latest dashboard release | v0.42.5 on 11 November 2026 | C | The changelog says that older releases may be left out when they had only internal changes. | [12] | | Time from first SuperPenguin page to the November 2026 check | 7 months | calc | Calculated: 12 April 2026 to 12 November 2026. | [9] | How to read this table. Grade C means the company or its founders said it on their own pages. Grade X means the value was calculated on this page. No row has a grade A or B source. ## Pricing SuperPenguin prices read from the company's pricing page on 12 November 2026 (grade C) | Plan | Per month, billed yearly (US$) | Per month, billed monthly (US$) | Limit | |---|---|---|---| | Free | | 0 | Up to US$2,000 managed spend; 1 workspace member | | Growth | | 30 | Up to US$5,000 managed spend; 3 workspace members | | Pro | | 200 | Up to US$20,000 managed spend; 10 workspace members | | Enterprise | | custom | More than US$20,000 managed spend; unlimited members | The page lists each price as US$ per month and does not say whether a yearly option exists. The plans are priced by the AI spend that SuperPenguin sees for the customer, not by seat. The page says that the SuperPenguin subscription itself never counts toward that limit. ## When did SuperPenguin (Carrot Labs) launch, and how did it grow? Timeline From two other tools in February 2026 to dashboard release 0.42.5 in November 2026. Every entry is the company's own page or a copy of it. - **5 February 2026** (grade C [5]): The archived Carrot Labs home page offers 3GPP Scout and an Earnings Excel plugin. The footer names Carrot Labs AI, Inc. - **6 March 2026** (grade C [3, 4]): Carrot Labs posts a model-tuning product on Y Combinator, in the Winter 2026 batch. The post lists two founders. - **12 April 2026** (grade C [9]): The privacy policy takes effect. It names SuperPenguin and Carrot Labs AI, Inc. - **26 June 2026** (grade C [13]): The oldest post on the SuperPenguin blog index is dated 26 June 2026. The index lists posts up to 18 August 2026. - **3 July 2026** (grade C [12]): The oldest changelog entry that is listed is v0.1.8. On 10 July the changelog adds a macOS desktop app. - **11 November 2026** (grade C [2, 12]): Release v0.42.5 is the current dashboard version. The pricing page lists Free, Growth, Pro and Enterprise. ## What could go wrong for SuperPenguin (Carrot Labs)? - **What is not verified.** No revenue figure was found, so no money result is shown. The pages read were the company's site, its customer stories, its Y Combinator page and an archived copy of its first site. No payment-processor profile, press interview or founder statement with a figure was found, because the search service was unavailable on 12 November 2026. The Channel3 and Freebuff pages are the company's own, and no independent source confirms that either customer pays. The sign-up and checkout pages were not opened, because the site's robots.txt closes them to automated readers. SuperPenguin is sold by Carrot Labs AI, Inc., which Y Combinator dates to 2026 and which showed other tools in February 2026. A reader could look for a revenue profile, a press piece or a founder post with a dated figure. - **Earlier products and the company's start.** The company existed before SuperPenguin. An archived copy of carrotlabs.ai from 5 February 2026 shows 3GPP Scout and an Earnings Excel plugin. On 6 March 2026 the company posted a model-tuning product on Y Combinator. The exact start of Carrot Labs AI, Inc. was not found; every source dates it to 2026 or later. The Y Combinator founder biography speaks of a continuous learning platform for AI agents, which matches the March 2026 product. - **Trust signals on the home page.** The strip of provider names (OpenAI, Anthropic, Gemini and others) lists services that SuperPenguin connects to, not customers. The Backed by Y Combinator logo agrees with the Y Combinator company page. The home page shows a sample dashboard with a US$22,478 bill and the customers Ridgepoint, Globex and Initech. It is a sample picture, not reported data. The home page prints one customer quotation, with a name, a role and the company Channel3. The same sentence is in the Channel3 case study. A second quotation, from Freebuff, is on the Customers page and in the Freebuff case study. - **Per-person coding data.** The coding view can show spend by person and by pull request. The terms tell the customer to follow monitoring and employment laws and not to use the estimates as the only basis for employment decisions. The documentation says a member sees only their own coding data unless an owner grants wider access. An owner can also turn on storage of full coding session transcripts for 90 days. The documentation says this is off by default and applies to the whole organization, with no per-person setting. The company's blog says that tracking AI use can be done without surveillance, and that the figures are for team trends, not individual rankings. The sources do not show any use against employees. ## Sources 1. SuperPenguin home page, read 12 November 2026 — https://superpenguin.ai/ (grade C). Read in full. Source for the product description, the FAQ, the provider strip, the sample dashboard and the customer quotations. 2. SuperPenguin pricing page, read 12 November 2026 — https://superpenguin.ai/pricing (grade C). Read in full. Source for the four plans, the managed-spend limits, the sign-up links and the billing questions. 3. Y Combinator company page: Carrot Labs — https://www.ycombinator.com/companies/carrot-labs (grade C). Read in full on 12 November 2026. Source for the batch, team size 2, founding year 2026, the founders' roles and the 14 providers. 4. Y Combinator launch post: Carrot Labs, custom models tuned to your tasks that never stop improving — https://www.ycombinator.com/launches/PdG-carrot-labs-custom-models-tuned-to-your-tasks-that-never-stop-improving (grade C). Read in full on 12 November 2026. The page data gives the creation time 2026-03-06. Source for the earlier model-tuning product. 5. Internet Archive copy of carrotlabs.ai, 5 February 2026 — http://web.archive.org/web/20260205014708/https://carrotlabs.ai/ (grade C). Read in full on 12 November 2026. Source for 3GPP Scout, the Earnings Excel plugin and the footer naming Carrot Labs AI, Inc. 6. Carrot Labs home page, read 12 November 2026 — https://carrotlabs.ai/ (grade C). Read in full. It names SuperPenguin as the company's current answer. 7. Internet Archive availability record for superpenguin.ai, 12 November 2026 — http://archive.org/wayback/available?url=superpenguin.ai×tamp=20260101 (grade C). Read in full. It lists a copy of 24 April 2026 as the closest to 1 January 2026. The copy itself could not be opened. 8. SuperPenguin terms of service, effective 14 August 2026 — https://superpenguin.ai/terms (grade C). Read in full. Source for Carrot Labs AI, Inc. as the provider, the fees at checkout and the monitoring-law clause. 9. SuperPenguin privacy policy, effective 12 April 2026, updated 22 October 2026 — https://superpenguin.ai/privacy (grade C). Read in full. Source for the first effective date, Stripe, Apollo.io and the default data collection. 10. SuperPenguin customer story: Channel3 — https://superpenguin.ai/customers/channel3 (grade C). Read in full. A company-written case study; no independent source confirms it. 11. SuperPenguin customer story: Freebuff — https://superpenguin.ai/customers/freebuff (grade C). Read in full. A company-written case study; no independent source confirms it. 12. SuperPenguin changelog, read 12 November 2026 — https://superpenguin.ai/changelog (grade C). Read the release list. Source for the version numbers and dates. 13. SuperPenguin blog index and Coding ROI page, read 12 November 2026 — https://superpenguin.ai/blog (grade C). Read the blog index in full and the Coding ROI page in full. Source for the post dates and the coding view. 14. SuperPenguin security page and sitemap, read 12 November 2026 — https://superpenguin.ai/security (grade C). Read in full. It says a SOC 2 Type II examination is under way; this is a claim of the company and is not shown as a tool. 15. Database of failed startups, local export — /method/#dead (grade C). Record pages read on the website on 12 November 2026. AI-written summaries, not independently verified. Source for the failed startups. 16. SuperPenguin documentation: workspace roles and AI Coding visibility; coding session transcripts — https://superpenguin.ai/docs/concepts/team-roles-and-visibility (grade C). Read both documentation pages in full on 12 November 2026. Source for what a member and an owner can see and for the transcript option. Contact: aiolum@livyden.com. Run SuperPenguin (Carrot Labs)? Send corrections or a verified figure to aiolum@livyden.com (subject: Correction: SuperPenguin (Carrot Labs)).