Littlebird

Littlebird is a context-first AI work assistant built around persistent memory rather than isolated chat sessions. The desktop app reads text and interface elements from the active window, listens during meetings for transcription and summaries, and can optionally connect to services such as email and calendar. That accumulated context powers chat, meeting preparation, cross-app recall, scheduled Routines and a quick overlay called Hummingbird. The company says it stores text rather than screenshots, which reduces the volume and visual sensitivity of what is retained but does not remove the underlying privacy stakes of continuous work-context capture.¹²³

The product now spans Mac and Windows, with iOS and Android companion apps; the iOS app can query context captured on desktop but does not itself capture new phone context.⁴⁵ Power plans add MCP access so Littlebird memory can be queried from compatible AI tools such as Claude and Cursor.⁶ TechCrunch’s March 2026 reporting broadly corroborates the architecture and described personalization becoming more useful after a few days of use, along with meeting notes, recurring routines and configurable app exclusions.³

Littlebird has a free Basic tier. Plus is $20 per month or $204 per year, Power is $50 per month or $510 per year, and Pro starts at $100 per month. Plus adds higher intelligence, enhanced memory, more usage credits, unlimited meeting notes, image generation and deeper research; Power adds roughly 2.5 times the Plus usage credits and MCP access. Team plans start at $17 per seat per month, while Enterprise adds options that include self-hosting.⁶⁷

Usage is credit-based on paid plans, so the subscription price is not always the full cost for heavy users. The current App Store listing also exposes usage-credit packs and in-app subscription prices, confirming that mobile purchases can differ from direct web billing.⁵ Buyers should compare expected chat, research and routine volume rather than assume the headline monthly tier is an unlimited-use price.

Independent exact-product coverage is still limited but not absent. TechCrunch reported on Littlebird in March 2026 after using it for several days, describing increasingly personalized prompts, meeting capture, Routines and configurable exclusions for apps and sensitive fields. The same article confirms that Littlebird stores cloud-hosted text context rather than screenshots and raised an $11 million seed round.³ The Information later referenced the same $11 million financing in its AI funding coverage.⁸

User evidence is positive but still small. The US App Store currently shows 4.3/5 from 48 ratings; its visible reviews include praise for persistent work context alongside complaints about occasional slowness or the need for stronger prompting.⁵ Product Hunt shows 5.0/5 from 10 reviews and summarizes recurring positives around reduced need to re-explain context, with requested improvements including deeper integrations, email handling, proactive task tracking and UI polish; one review mentions occasional hallucinations.⁹ These are useful directional signals, not large-sample reliability evidence.

Littlebird’s privacy posture is unusually important because the product is designed to observe work continuously. The company states that memory is encrypted and stored on AWS, with AES-256 at rest, TLS 1.3 in transit and AWS KMS for key management; it also states SOC 2 certification, GDPR- and CCPA-aligned handling, deletion controls and a policy of not selling personal data or using it to train AI models.¹⁰¹¹ The standard consumer architecture is not local-only. Littlebird explicitly says users who require local processing should not use the ordinary individual setup; self-hosting is reserved for Enterprise.⁷

The Privacy Policy says collected information can include screen content, browser activity, supported-app interactions, chat messages, task and relationship data and account information.¹⁰ The Terms describe the service as monitoring screens and interactions to help manage tasks, projects and information.¹² That breadth means app exclusions, capture controls, retention/deletion settings and employer or client confidentiality rules should be treated as deployment requirements, not optional housekeeping.

Littlebird is best suited to knowledge workers whose main AI friction is context reconstruction: people who move across documents, email, meetings, browsers and project tools and repeatedly have to explain what they are doing before an assistant can help. The strongest fit is for users comfortable trading some data centralization for a continuously updated personal work memory.

It is less suitable for environments where confidential work cannot be sent to cloud infrastructure, where continuous context capture conflicts with internal policy, or where a large independent reliability record is required before adoption. Teams should also distinguish the desktop capture model from the mobile companion experience, because the current iOS app retrieves desktop context rather than independently building the same memory on-device.⁵