Melissa Blake

Senior Editor · AI Companion Platforms · Amsterdam

I write about AI companion apps and character-chat platforms as products people actually use. The PixelChat AI page on this site is mine, approach, sourcing, corrections and all.

Short version

Senior editor covering AI companion apps and character-chat platforms since 2022, when the category first started separating itself from general-purpose chatbots. My beat is consumer-facing: the products people sign up to, chat with, and decide whether to pay a subscription for. This site is where the PixelChat-specific writing lives, with cross-reference to SpicyChat where the shared account system materially affects a reader's decision. Wider category reading informs comparisons on the main page, not standalone coverage here.

The beat

Consumer AI companion apps, character-chat platforms, and adjacent tools. PixelChat is the focus of this site, with attention on the ways it connects to SpicyChat because both products belong to the same publisher (NextDay AI). My wider coverage window includes Character AI, Replika, Chai, Janitor AI, JanitorAI, Talkie, Kindroid, PolyBuzz, Candy AI, SpicyChat, Chub and PepHop. That range gives comparisons on the main PixelChat page a wider point of reference than a PixelChat-only view would.

How I approach a review

Live product, real account, over a long enough window that second-week behaviour shows up rather than just the onboarding pitch. For PixelChat specifically, that means observing free-tier ad frequency, character discovery through the tag filter and Trending sort, memory drift over long chats (Get A Taste 4K context fills up faster than most reviews warn), the client-side refusal pattern Google Play reviewers describe when a scene turns explicit, and the tier-gated features documented at docs.pixelchat.ai. Paid-tier detail is cross-checked against pixelchat.ai/subscribe rather than repeatedly cycled through billing.

Claims about pricing, feature gating, content-policy behaviour, or the PixelChat–SpicyChat relationship get verified against current sources before publication rather than carried forward from an older draft. Where a claim rests on published research (loneliness effects, adolescent risk, engagement patterns across the wider category), the source is cited inline with sample size and publication venue. Testing detail is documented on the methodology page.

What I don't do

No sponsored coverage. No advance sight for platforms of critical material about themselves. No coverage of a product I hold a financial interest in beyond the disclosed affiliate revenue detailed on the editorial policy page. No fabricated first-person measurements: personal-testing statements that would require a specific numeric result (a wait time, a token count, a refusal rate) get attributed to named sources such as documentation, dated user reviews or third-party reporting rather than invented. If any of that changes, editorial policy is where the change gets logged first.

Contact

Primary channel is LinkedIn: linkedin.com/in/melissablake1. Corrections, review requests and press queries route to the same address. Corrections get a seven-working-day turnaround. Press queries move around active editorial work on a best-effort basis.

Location note

Amsterdam-based. Time-of-day observations in the PixelChat coverage (evening ad-frequency shifts on free tier, for example) anchor to Central European Time by default. Where a timing claim would land differently across time zones, it is expressed in UTC.

Bylines on this site

Profile last reviewed: October 6, 2026

Read next