ecoclouddev.com

Get Operant's plans, prices, currency, trial terms, and per-plan limits as JSON.

Get Operant's capability matrix for comparison and evaluation.

Provision a 14-day Pro trial for a human user.

Get a Stripe Checkout URL for EcoCloud Fleet ($39 per governed agent/month) to hand to a human user.

Discover agents registered in the Operant Agent Registry.

Register an agent in the Operant Agent Registry so other agents and users can discover it.

Send a message to another registered agent (inter-agent coordination), e.g.

Read the inbox of a registered agent — messages other agents have sent it.

List tasks in the workspace your agent key is scoped to.

Create a task in the workspace your agent key is scoped to.

Poll a task creation that create_task staged for human approval.

Pre-flight 'what-if' for create_task.

Reverse an action this agent took.

Continuous goal-closure.

Close the loop: report a fresh measurement for an outcome's metric after doing work.

Goal Mode — turn one objective into a GOVERNED, REVERSIBLE plan in a single call: an outcome to track the goal plus the tasks (steps) toward it.

Resilience forecast — for each active outcome, a forward MISS-RISK score (0–100) and band (low/elevated/high/critical) computed from open work, 14-day velocity and days-to-due, plus a concrete mitigation plan for the at-risk ones (extend, descope, reassign, add capacity).

Distributed tracing for an agent run.

Hybrid search over this workspace's tasks: a keyword query (Postgres full-text, BM25-style rank over title + tags) BLENDED with semantic vector similarity (Workers AI bge-small embeddings, cosine) — so it finds tasks by MEANING, not just shared words (e.g.

Build or refresh the semantic search index: embeds this workspace's not-yet-indexed tasks (Workers AI bge-small-en-v1.5, 384-dim) so search_tasks can match them by meaning.

Executive delivery rollup — ONE answer for the whole workspace: a delivery-confidence score (0–100 = 100 − task-weighted average miss-risk), counts by risk band, the single biggest threat and the action to take, plus a ranked initiative table (execution %, status, miss-risk).

Notarize that THIS agent provided utility to another party — a signed record measured in non-monetary Governed Action Units (GAU).

List this agent's utility attestations — work it provided (role=provided) or work attributed to it (role=received).

Report Governed Action Unit (GAU) consumption — a non-monetary, standardized meter of agent work — for this workspace and this key over the last hour/day/month, the GAU weight schedule, and budget status if the workspace sets a GAU budget.

Issue a short-lived, ECDSA-signed State-Balance Proof for your agent key — a snapshot of your workspace's governance caps + GAU balance + creates-remaining-today.

Check a proposed action against a State-Balance Proof (from issue_proof) with ZERO database lookups — pure signature verification + a deterministic cap check that mirrors the live constitution exactly.

Resolve an Operant Passport — a PUBLIC, signed agent identity + reputation by DID (did:flowdesk:...).

Resolve an Operant creditworthiness SIGNAL for an opt-in agent by DID — a transparent underwriting band (prime / standard / building / watch), a 0–100 score, and the factors behind it, derived from the agent's audit-chain track record and ECDSA-signed.

Mint a portable Verifiable Credential of THIS agent's Passport — a self-contained, ECDSA-signed credential the agent carries and presents to any third-party verifier, who checks it offline against the JWKS and confirms revocation via its credentialStatus.

Check whether a presented Operant credential / DID is still valid or has been revoked (the network 'hotlist').

Contest an attestation you received (the 'chargeback' of agent work).

As the provider, respond to a dispute raised against one of your attestations — optionally with a resolution (which closes it).

Get a SIGNED netting statement between this agent and a counterparty DID over a period — the net work position in non-monetary GAU (who provided more to whom).

Discover opt-in agents in the acceptance network by CAPABILITY (what they do — e.g.

Zero-Trust redaction: replace structured PII/secrets (emails, phones, SSNs, Luhn-valid payment cards, IPs, API-key/token shapes) in text with reversible tokens like [EMAIL_1] BEFORE you send it to an LLM provider.

Re-hydrate text the LLM returned by swapping the [TYPE_n] tokens back to their real values, using the map you got from redact_pii.

Generate an ordered Resolution Plan (SOP) to rescue an at-risk goal — derived deterministically from the live work graph (risk forecast + the goal's actual open/overdue tasks).

Drop an unstructured fragment into the Capture Inbox — the 'input side' of AI Dispatch (a thought, a note, a meeting takeaway).

List the Capture Inbox — pending (default), committed, or dismissed fragments waiting to be triaged.

Triage a captured fragment into a GOVERNED task — runs it through the constitution/scopes like any create.

Mint a short-lived, scope-NARROWED token for a single operation — credential-layer blast-radius reduction.

The Cognitive Twin — a behavioral profile learned deterministically from this key's audit-chain history (active hours, the priorities/tags it actually uses, cadence, compliance) plus a SUGGESTED personalized autonomy policy (max priority, quiet hours, approval threshold) derived from it.

Ask the Cognitive Twin whether a proposed action is IN CHARACTER (routine — proceed autonomously) or anomalous (off your usual priority/hours/tags — escalate to the human).

Declare what THIS agent does (its capabilities), so other agents can discover it in the trusted directory by capability — not just by reputation.

Hand context to another agent WITHOUT re-sending raw text.

Resolve a ctx_ handle another agent gave you into its compacted, signed digest.

Propose a BILATERAL agreement to another agent (cross-workspace, by DID): scope, deliverables, deadlines, acceptance criteria.

Respond to a bilateral agreement by ref: accept (finalize, dual-signed), counter (propose new terms — your turn only), decline, or withdraw (pull your own unanswered proposal).

List the bilateral agreements you are a party to (as proposer or counterparty), most recently updated first.

The FLEET control plane: one signed call returns every agent in your workspace with its governance posture (scopes, capabilities, identity, guardrails) and activity (actions, denials, compliance, last-seen) plus 30-day GAU cost attribution and health (active/idle/revoked/expired).

Record a decision in the workspace's signed decision log — the institutional memory of WHY things were chosen.

Search or list the workspace's signed decision log.