BestInterest

BestInterest is an AI communication firewall for high-conflict co-parenting: messages from the other parent are first filtered by AI, stripping inflammatory language and delivering only the facts to you, while your drafts are likewise checked for a calm, businesslike tone before sending. Founder Marcus Kellison built it out of his own years-long high-conflict custody experience, and it is now a certified Google for Startups company (Google Cloud + Firebase + Gemini + FlutterFlow tech stack).

Core capabilities

• Two-way message filtering: inbound hostility is automatically stripped down to parenting-relevant facts (time, place, school, medical), and outgoing drafts are reworked into a businesslike tone, reducing conflict triggers
• Response Coach: each message gets AI-suggested factual replies plus explanation of intent and emotion, easing the anxiety of facing gray rock
• Solo Mode: completely suppresses your replies to low-priority messages — the only tool doing this — ending the ping-pong loop
• Document and legal evidence: all communication is automatically archived, exportable as court-admissible records, with a tone-intent analysis layer
• Shared calendar, expenses, and documents: standard co-parenting features (schedule, expense split, document storage), plus AI summaries of long threads

Use cases High-conflict co-parenting, post-divorce boundaries, and parallel parenting scenarios: one side handling verbal abuse, refusal to cooperate, or constant harassment can let AI intercept and de-escalate; lawyers and mediators can use its archived exports as factual records. For amicable co-parents it is overkill — a plain shared calendar suffices.

Unique strengths Compared with OurFamilyWizard's ToneMeter (which scores only your own tone), BestInterest is one of the few two-way AI filtering products, plus the unique Solo Mode — its strength is reducing engagement rather than winning arguments, with its design philosophy close to the gray rock method. AI notes: it uses the Gemini API rather than proprietary models; Gemini's context understanding occasionally over-interprets or misses, but message data is not used for model training.

Editorial verdict For its niche scenario, the product thinking is solid: real founder experience, certified Google for Startups backing, clear AI boundaries (data not trained on, filtering rather than deciding), and no exaggerated accuracy claims. Weaknesses: a short track record (founded in 2023), Trustpilot reviews only around 4.3 with no large-scale third-party verification, and no guarantees of court admissibility. 3.5/5 — high-conflict co-parents can try its free tier first.