# Alkera

Aiolum story: https://aiolum.com/story/alkera/ · added 2026-11-12 · updated 2026-11-12

**In short.** Alkera (AI agent and workspace for data engineering, data analysis and data science, USA): Live with paid plans, November 2026. Evidence grade E: no revenue figure was checked. Team: 3 founders; the Y Combinator page lists a team size of 3. Launched July 2026; still running.

AI agent and workspace for data engineering, data analysis and data science · USA · AI-native · team of 2 to 10 · still running · group: Documents, data and slides

- Outcome: Live with paid plans, November 2026 (evidence grade of the headline number: E; no revenue figure)
- Launch: July 2026; months from launch to the outcome: 4
- Team: 3 founders; the Y Combinator page lists a team size of 3
- Outside funding: Y Combinator, Fall 2026 batch; amount not published
- Main factor: Domain expertise
- Meters, 1 to 4: evidence strength 1, copyable pattern 1, capital needed 2
- Tags: AI-native, Small team, Funded, Running, B2B, Data

Links:

- Visit alkera.ai: https://www.alkera.ai/

**The product.** A web workspace where a team and an AI agent do data work together: pipelines, analysis and model training. The agent reads the company's data tools, traces which tables and dashboards a change would touch, and runs code in notebooks. The company sells Free, Plus and Pro plans and a custom Enterprise plan.

**Why it worked.** The three founders say they worked with data teams at several companies before they started Alkera. The Y Combinator page and the launch post list earlier work at Verition Fund Management, ByteDance, Ramp and Hudson River Trading. Their launch post says coding agents can write data code that runs, but knowing it is correct is the hard part.

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

Alkera is an agent for data engineering, the work of building and repairing the pipelines that move and transform company data. It also covers analysis and data science. Its pages describe a knowledge base of the company's definitions and column-level lineage, a map of which column feeds which table and dashboard. The agent uses the map to check what a change would break.

Rick Gao, Tony Li and Andrew Tran are named as founders on the Y Combinator company page, which says "Founded: 2026" and "Team Size: 3". The company's launch post on that site is dated 17 July 2026. The pricing page lists a free plan, Plus at US$50 a month and Pro at US$200 a month.

No revenue figure, customer count or funding amount was found, so this story has grade E. The company's terms set monthly billing, and its privacy page names Stripe as the payment processor. No source states that a customer has paid.

## Why did Alkera work?

### The founders had worked with data teams before they started

The Y Combinator page says Rick Gao built financial models at a fund and Tony Li built agent tooling at ByteDance. It says Andrew Tran was a software engineering intern at Ramp and HRT, "working closely with data teams". These are the founders' own descriptions of their past work. No source shows that this work led to the product or to any customer.

### The agent maps what a change touches before it edits

The launch post says the agent traces column-level lineage from the warehouse through dbt models to the dashboard. It says the agent lists every downstream table and dashboard before it changes anything. This is the company's description. No source here tests it, and the benchmark result is the company's own.

### It offers a free plan, two paid plans and open-source code

The pricing page offers a free plan and two paid plans, Plus and Pro. The company also publishes Databench under the Apache-2.0 licence. It describes Databench as an open-source subset of the product. No source says how many users come from either route.

## How it was built

These details come from the company's own pages and its GitHub page. They are grade C.

- The agent works with model providers. The company's Databench page says a model gateway "calls Anthropic or OpenAI with your keys". The enterprise page lists Claude, Codex, Gemini, DeepSeek and Kimi.
- The workspace runs in the browser. The documentation says the agent and the notebooks run on managed machines, and that IDE extensions and a command-line tool are "coming soon".
- The notebooks are SQL and Python notebooks. The Product Hunt post says they are based on marimo.
- Payments run on Stripe. The website uses Google Analytics and Cloudflare Turnstile, as the privacy policy says.

Sources: [7, 14, 10, 9, 8, 4]

## Who tried this before Alkera, and why did they fail?

Three companies in our database of failed startups tried to make data work easier for people who are not full-time data engineers. All three ended. The values below are from that database, whose pages call them AI-assisted summaries, 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.

### DataNitro (2012 to 2015; funding US$120K; cause: Competition)

- Promise: Python scripts that run inside Excel, so analysts could move spreadsheet work to Python without leaving Excel.
- Why it died: The record says the product was limited by Excel's design and that Excel later added Python itself.
- Record: 

### Dashblock (2019 to 2022; funding US$3.0M; cause: Technology)

- Promise: Turn any web page into an API with no code, so businesses could use web data without programming.
- Why it died: The record says the tool depended on web page structures that kept changing, which meant constant repair work.
- Record: 

### Scratch Data (2021 to 2024; funding US$5.0M; cause: Competition)

- Promise: An on-demand big data platform, so companies could run data stacks without in-house experts.
- Why it died: The record says it could not stand out from the big cloud providers, which offered integrated data services.
- Record: 

## What changed

Each of the three offered a simpler way to do data work. Each sat on a platform that someone else owned: Excel, the web or the cloud. The records say Excel added Python itself, web pages kept changing, and the big cloud providers offered integrated data services. Alkera also sits on top of other vendors, because its pages list connections to warehouses, dbt and Airflow.

One thing differs from the three records. The central part of Alkera is an AI agent that takes tasks in plain language. Its pages say the agent checks a change before it makes it. The sources do not show whether this holds up on real company data, and no revenue figure exists.

## Can one person copy Alkera now?

**Verdict: The idea can be studied. The result is unknown.**

Three founders with data backgrounds built an agent for one job: safe changes to data pipelines. They published prices and an open-source part. No source shows that any customer has paid, so a reader cannot judge the return.

1. Pick one job that your own team knows well and where mistakes are costly, such as changes to data pipelines.
2. Show a check before the agent acts, for example a list of everything downstream that a change would touch.
3. Publish a price list from the start, with a free plan and two paid plans.
4. Publish an open-source part of the product, so that engineers can try it without a sales call.

**What is still open.** Other jobs where an agent must prove its change is safe before it acts: database migrations, financial reports, and access rules for company data.

## How much revenue does Alkera make? The numbers, graded

A three-founder company from 2026 with a published price list and billing terms. Revenue, customers and the funding amount are not public

| Fact | Value | Grade | Note | Sources |
|---|---|---|---|---|
| Status, November 2026 | live, with Free, Plus, Pro and Enterprise plans | C | The pricing page lists the plans with "Get Plus" and "Get Pro" buttons. The documentation says Alkera runs in the browser at app.alkera.ai. | [2, 9] |
| Founded | 2026 | C | The Y Combinator company page says "Founded: 2026", "Batch: Fall 2026" and "Team Size: 3". It lists the company as Active. | [5] |
| Founders | Rick Gao (CEO), Tony Li (COO), Andrew Tran (CTO) | C | The Y Combinator page lists the three as founders. The launch post names the same three under "Our team". | [5, 6] |
| Earlier name or product | none found | C | The home page, terms, privacy page and Y Combinator page show only the name Alkera and the company "Alkera AI, Inc.". The earliest archived copy of the home page is from 8 October 2026. | [1, 3, 4, 5, 13] |
| Launch post | 17 July 2026 | C | The launch post on the Y Combinator site carries the creation date 17 July 2026. It names the product as an agent for data engineering, analysis and science. | [6] |
| Price list | Plus US$50 and Pro US$200 a month | C | The pricing page. The page lists no yearly price. Its Enterprise plan has a custom price. The documentation's billing page gives Pro as "$250 / month". | [2, 15] |
| How customers pay | monthly, card payments through Stripe | C | The terms say "Customer authorizes Provider to bill and charge Customer's payment method on file Monthly". The privacy page says "Payments are processed by our third-party payment processor, Stripe." | [3, 4] |
| Open-source workspace | Databench, Apache-2.0 licence, 13 stars and 1 fork on 12 November 2026 | C | The company's GitHub repository. The company describes it as an open-source subset of the Alkera product. | [7, 8] |
| Product Hunt launch of Databench | 182 points, 475 followers, launched about a month before 12 November 2026 | C | The Product Hunt page shows "Launched 1mo ago", "182 points" and "475 followers". Rick Gao, Andrew Tran and Tony Li are marked as makers. | [8] |
| Company's benchmark claim | 83.28% against 68.49% for Claude Code | C | The home page says DataAgentBench, average pass rate over 5 runs, July 2026. It is the company's own result. | [1] |
| Difference of the two benchmark results | 14.79 percentage points | calc | Calculated: 83.28 minus 68.49. The home page rounds it to "15 points". | [1] |
| Outside funding | Y Combinator, Fall 2026 batch; amount not published | C | The home page shows "Backed by" with the Y Combinator logo. The Y Combinator page lists the batch. No downloaded page gives an amount. | [1, 5] |

How to read this table: a grade C value is a statement on the company's own pages or on its Y Combinator profile. A "calc" value was computed on this page. No value is grade A or B, and no revenue amount exists in the sources.

## Pricing

Alkera plans in US dollars on 12 November 2026, from the company's pricing page (grade C)

| Plan | Per month, billed yearly (US$) | Per month, billed monthly (US$) | Limit |
|---|---|---|---|
| Free | 0 | 0 | 10 GB storage; limited usage |
| Plus |  | 50 | 1 TB storage; extra usage; access to all models; marked Most popular |
| Pro |  | 200 | 5 TB storage; significantly increased usage; dedicated support; early access to new features |
| Enterprise |  |  | Custom price; pooled usage and storage; SSO; audit logs; VPC and on-premises |

The page gives one price per month and no yearly price. The documentation's billing page gives Pro as "$250 / month"; this table uses the pricing page. Plus and Pro add storage and usage over the free plan, and Pro adds dedicated support and early access. Enterprise has no price on the page and lists single sign-on, audit logs and on-premises deployment.

## When did Alkera launch, and how did it grow? Timeline

From terms of service in June 2026 to a live price list in November 2026. Each entry carries its grade.

- **17 June 2026** (grade C [3]): The terms of service are last updated. They name Alkera AI, Inc. as the provider and say that customers pay fees by plan and usage.
- **17 July 2026** (grade C [6]): The launch post of the company appears on the Y Combinator site. It says the product is number 1 on UC Berkeley's DataAgentBench.
- **29 July 2026** (grade C [4]): The privacy policy is last updated. It names Stripe as the payment processor.
- **8 October 2026** (grade C [13]): The earliest copy of the home page in the Internet Archive.
- **25 October 2026** (grade C [13]): The earliest copy of the pricing page in the Internet Archive. A later copy of 6 November 2026 exists.
- **October 2026** (grade C [8]): Databench, the open-source workspace, launches on Product Hunt. The page reads "Launched 1mo ago" on 12 November.

## What could go wrong for Alkera?

- **What is not verified.** No revenue figure, customer count or funding amount was found. We looked at the pricing page, the terms, the privacy policy, the Y Combinator pages and the Product Hunt page. The paid plans are on sale: the pricing page has prices and sign-up buttons, and the terms set monthly billing. The privacy page names Stripe as the payment processor. No source states that a customer has paid. The "number 1 on DataAgentBench" claim is the company's; the leaderboard table on the benchmark page was empty in the download. A reader could look for a payment-processor profile or a press piece.
- **Home page logos, counters and chat examples.** The home page shows one "Backed by" logo with the text "Y Combinator" and no investment amount. Its wall of vendor names, such as Snowflake and dbt, lists tools that Alkera connects to, not customers. The chat examples show people named Priya and Marcus in sample conversations. The bar chart of 83.28% against 68.49% is the company's own result. The enterprise page says "GDPR compliant" and "SOC 2 Type II, ISO 27001 & HIPAA in progress"; these are the company's statements.
- **Pages that describe the product differently.** The Y Combinator page says the agent works in an IDE or a command-line tool. The documentation says it runs in the browser and that IDE extensions and a command-line tool are coming soon. The Y Combinator page lists the batch as Fall 2026, while its launch post is dated 17 July 2026. The pricing page gives Pro as US$200 a month, while the documentation's billing page gives "$250 / month". This story uses each statement as its source gives it.
- **AI-written articles on the company's site.** The company's Perspectives pages say "Powered by The Prompting Company" and "This site contains AI-generated content, which may have errors". One article cites results from a hedge fund customer without naming it. No downloaded page confirms that case study, so this story does not use it.

## Sources

1. Alkera, home page — https://www.alkera.ai/ (grade C). Read in full on 12 November 2026. Source for the product description, the vendor wall, the benchmark claim and the Backed by logo.
2. Alkera, Pricing — https://www.alkera.ai/pricing (grade C). Read in full on 12 November 2026. Source for the four plans and the prices.
3. Alkera AI, Inc., Terms of Service, last updated 17 June 2026 — https://www.alkera.ai/tos (grade C). Read in full on 12 November 2026. Source for the provider, the fees and the monthly billing.
4. Alkera AI, Inc., Privacy Policy, last updated 29 July 2026 — https://www.alkera.ai/privacy-policy (grade C). Read the sections on payments and analytics on 12 November 2026. Source for Stripe, Google Analytics and Cloudflare Turnstile. It also has a street address, which this story does not repeat.
5. Y Combinator, Alkera AI company page — https://www.ycombinator.com/companies/alkera-ai (grade C). Read in full on 12 November 2026. Source for the founding year, batch, team size and the three founders.
6. Y Combinator, launch post: Alkera - The data agent that you can trust — https://www.ycombinator.com/launches/Rcy-alkera-the-data-agent-that-you-can-trust (grade C). Read in full on 12 November 2026. Source for the 17 July 2026 date, the product claims and the team descriptions.
7. GitHub, AlkeraAI/Databench — https://github.com/AlkeraAI/Databench (grade C). Read the repository front page on 12 November 2026. Source for the licence, the stars and the forks.
8. Product Hunt, Databench by Alkera — https://www.producthunt.com/products/alkera (grade C). Read in full on 12 November 2026. Source for the points, followers, launch time and the maker's description.
9. Alkera documentation, home page — https://docs.alkera.ai/ (grade C). Read in full on 12 November 2026. Source for the browser app and for IDE extensions being coming soon.
10. Alkera, Enterprise page — https://www.alkera.ai/enterprise (grade C). Read in full on 12 November 2026. Source for the models listed and the enterprise features.
11. DataAgentBench, overview and leaderboard — https://ucbepic.github.io/DataAgentBench/ (grade C). Read in full on 12 November 2026. The leaderboard table was empty in the download, so Alkera's rank could not be checked.
12. Alkera, Perspectives article: How Alkera holds up at sequencing scale — https://www.alkera.ai/perspectives/blog/explainer/how-alkera-holds-up-at-sequencing-scale-verifying-intermediate-results (grade C). Read in full on 12 November 2026. Source for the AI-generated content notice and the unnamed hedge fund figures.
13. Internet Archive index of alkera.ai and its pricing page — https://web.archive.org/cdx/search/cdx?url=alkera.ai/pricing&limit=5&output=txt (grade C). Read the capture list on 12 November 2026. Earliest copy of the home page 8 October 2026 and of the pricing page 25 October 2026.
14. Alkera, Databench page — https://www.alkera.ai/databench (grade C). Read in full on 12 November 2026. Source for the model gateway and the agent harness of Databench.
15. Alkera documentation, Billing & usage — https://docs.alkera.ai/organization/billing (grade C). Read in full on 12 November 2026. Source for the plan table that gives Pro as $250 a month, and for credit top-ups paid through Stripe.
16. 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 predecessors.

Contact: aiolum@livyden.com. Run Alkera? Send corrections or a verified figure to aiolum@livyden.com (subject: Correction: Alkera).