Tech

Wabi, the AI agent that builds mini apps, now valued at $1.25B after latest round

TechCrunch·Oct 29, 2026 · 9:35 AM CT
Read the original at TechCrunch
Readthrough summary

Wabi, which now calls itself a "personal agent," said Thursday it has raised $55 million at a $1.25 billion valuation.

The company was founded by Eugenia Kuyda, who previously founded the AI companion startup Replika. Wabi originally let anyone build apps from prompts, but last month it pivoted to become an AI messenger, dubbed Wabi 2.0, that builds the interfaces a user needs as part of its role as a personal agent.

Wabi suggests mini apps for meal planning, tracking local restaurants, date-night planning, fitness plans and custom trivia games. TechCrunch notes the funding comes amid a hot market for AI agents and companions.

Key points

  • Raised $55M at $1.25B valuation
  • Founded by Replika creator Eugenia Kuyda

Summary written by Readthrough from the original article. The full story, with photos and any updates, is at TechCrunch.

Read the original at TechCrunch

More Tech

Cloudflare Blog·

Cloudflare DDoS Threat Report 2026 Q3: More swarms, fewer giants

Cloudflare's Q3 2026 DDoS report finds short, small attacks (under 10 minutes and 1 Gbps) surged 208% while hyper-volumetric attacks over 1 Tbps fell nearly 69% from Q2, possibly after law enforcement took down four big IoT botnets. Attackers also dropped CLDAP floods almost entirely and shifted to DNS amplification, up 215%.

9to5Google·

Google had a very good, very bad Q3

Alphabet's third quarter brought a 25% revenue jump, $26 billion in net income and a cloud business that doubled revenue to $30 billion, yet the stock fell as much as 11% after hours because ad revenue missed expectations. 9to5Google says investors worry search ads aren't keeping up with users shifting to AI.

Amazon News·

Alexa for Shopping adds 100+ new capabilities to make shopping even faster and easier

Amazon says Alexa for Shopping has added more than 100 new capabilities, letting customers manage lists, registries, Subscribe & Save, payments and memberships just by asking. Over 400 million customers used it in the past 12 months, and use of its agentic features has grown more than eightfold since January.

ZDNet·

Mistral’s Le Chonk is now a top open model for coding, ranking alongside Codex GPT-6.1 Terra

French lab Mistral has released the weights for Mistral Large 4, a trillion-parameter model nicknamed "Le Chonk." Its score on the Terminal-Bench 4.0 agentic coding test jumped from 28% in preview to 41%, edging GLM-5.3, and its overall index score of 48 ties Codex GPT-6.1 Terra, ZDNet reports.