flmnt sits behind whatever agent you already run. Start free on a single project, scale to a whole team, or bring it into your own cloud. Limits aren't hard caps — you never hit a wall mid-session.
One agent, one project. Enough memory to feel the difference on a real codebase.
The everyday plan for builders shipping with agents daily. The full memory model, one flat price.
Shared memory across a whole team — roles, audit trail, and the priority lane when something's on fire.
A dedicated single-tenant deployment with private network peering, signed paperwork, and a custom SLA.
Prices in USD. Annual plans billed once up front (10 months for 12). All plans include the full memory model — keyframes, decisions, mistake tracking, and causal retrieval.
The full feature set across all four plans. Metered limits are starting allotments — overages bill at the rates in the next section, never as a hard stop.
† Response-time SLAs are measured during business hours — 9–5 PT, Monday–Friday.
Plan limits are starting allotments, not walls. When you exceed one, your agent keeps writing and reading — the extra usage simply accrues on your next invoice at a flat per-unit rate. You only pay for what you actually store and query.
Say you're on Pro and store 30,000 memories in a busy month against the 25,000 included. That's 5,000 over — five units at $2.00 — about $10 in overage. Memory stays cheap because reads and retrievals are never metered.
Overage is a flat $2.00 per 1,000 memories across paid plans. Need a bigger included allotment or predictable spend? Enterprise sizes it to you.
We stand up a dedicated, single-tenant flmnt deployment with private network peering to your environment, sign your DPA and MSA, and provide a named CSM with a custom SLA. Talk to us about residency, security review, and in-tenant inference.