The autonomous lifecycle agent for Four.meme tokens on BNB Chain. Trust grades are deterministic (pure on-chain rules, zero LLM). Operational decisions go through an LLM gateway and every call is Merkle-committed on-chain.
1,573 DGrid LLM calls + 518 MYX decisions committed to on-chain Merkle roots · 6 attestation txs · $AUNT launched by the agent on Apr 20
Paste an address below. Same endpoint the SDK hits. Same answer the agent uses. Zero LLM in the trust path — the rules that fired are shown below the grade.
Every trust grade is computed from raw on-chain metrics with a full why[]rule trace. No LLM is used to decide if a token is safe — anyone can reproduce any grade from the raw data. The agent's operational LLM calls (narrative picks, post content, hedge signals) go through a gateway that commits every call to an on-chain Merkle root.
Five tiers — graduated, watch, healthy, at_risk, observed — computed from on-chain metrics with a full rule trace. Pair-aware graduation targets sourced live from Four.meme.
Declare defensive thresholds per token. On critical verdict, FOUR-LIFE halts content posts, fires a webhook, and emits a short-hedge signal — all deterministically.
Signed HMAC webhooks for every tier and protection transition. Python + TypeScript SDKs. Telegram + Discord fan-out. Build on top in minutes.
A Four.meme token enters on the left. The autonomous agent runs a three-phase lifecycle loop. Every trust grade is deterministic (no LLM); every operational LLM call is committed to an on-chain Merkle root. Four consumer surfaces read the same primitives.
Read left→right: a token enters, the agent grades it with pure rules while logging every LLM reasoning step on-chain, and four surfaces consume the same primitives.
Every trust grade is computed from raw on-chain metrics — holders, whale concentration, buy/sell ratio, curve progress, contract risk — against deterministic thresholds. The rules that fired travel with the verdict so anyone can reproduce the grade.
Every grade is written to a time-series store. Tier changes + keepalive snapshots. Query the history, compute diffs, export NDJSON for backfills.
On every transition, fire signed HMAC webhooks and fan out to Telegram + Discord. Subscribers verify, react, and halt trading in real time.
Parallel SDKs for Python and TypeScript, covering every endpoint. Sync or async. Zero config. Write custom agents, alert bots, or treasury guards in minutes.
pip install four-lifePyPInpm install @gudman/four-life-sdknpmcurl https://four-life.gudman.xyz/api/token/<addr>/badgeHTTPTraders read the radar, creators embed the badge, operators subscribe to webhooks, and bots trigger off tier transitions. Same deterministic grade across every surface.
Every live Four.meme token, ranked deterministically. Filter by confidence, sort by graduation odds. Click for the full rule trace.
React to tier changes and protection transitions in real time. HMAC-signed, retried, verified. Built for autonomous trading bots.
One line of HTML drops a live Certified badge on your Discord, site, or bio. Auto-refresh. No build step. No tracking.
The badge belongs next to every Four.meme token address. A composable trust primitive — embed the widget, consume the API, fork the SDKs.
Drop-in platform primitive. Creator widget. Trader radar. Autonomous defender. Zero trust cost. Open source.