Melaya — Build AI agents for any job. Agentic platform for research, ops, outreach, reporting — and the only one where agents can actually trade.

// USE CASE · OPERATIONS

Operations agents that watch every system,and page you before it breaks.

Most ops leads spend the morning stitching dashboards together by hand and the afternoon explaining a stale number to leadership. Melaya lets you build an operations crew that reconciles positions, watches API health, drafts the daily brief, and triages incidents into Jira or Linear. The crew handles the polling and the paperwork, you keep the judgment calls, the audit log captures every write.

See the pipelines ↓
01
// What breaks today

The status quo costs more than the agent does.

Three pains every sales and BD team hits weekly. Each one is what your reps actually complain about, not what a feature page would call them.

  1. 01

    Ops leads spend two hours every morning copying numbers from five dashboards into one slide deck, and the deck is stale by the time leadership opens it.

  2. 02

    A degraded data feed sits unnoticed for six hours because no one owns the freshness check, then the trading desk discovers it mid-session.

  3. 03

    An incident lands in three Slack channels, a Jira ticket, and a Telegram thread with no single timeline, and the post-mortem takes a week to reconstruct.

02
// Pipelines you can build

Compose. Approve. Replay.

Every pipeline below is a shape you wire on the canvas using the crew and tools further down. Not a feature we ship for you, a pattern you configure.

P01

Reconcile positions across exchanges

On cron, TradingOperations calls melaya_get_positions and melaya_get_account_balance per venue, computes net exposure and margin utilization, and flags any settlement gap. Static context holds the per-venue risk limits so scoring stays consistent across runs.

P02

Watch data freshness on every feed

DataOperations probes API endpoints via http_request and checks melaya_ohlcv timestamps against the documented SLA matrix. The rag_retrieve tool pulls the SLA doc on demand so freshness thresholds match the runbook the team already maintains.

P03

Triage incidents into one timeline

On a flagged anomaly, IncidentManager classifies SEV1, SEV2, or SEV3, opens a Jira or Linear ticket, and pages on-call via slack_post_text and telegram_send_message. HITL gate blocks every outbound page until severity and owner are confirmed.

P04

Ship the morning operations brief

ReportingAnalyst pulls SQL aggregates with row count and freshness flags, OperationsSynth ranks them into the four-section brief, then slack_post_text drops it in the leadership channel. Cross-run memory keeps yesterday's open items in the brief until they close.

P05

Track vendor and filing deadlines

ComplianceOperations reads gcal_list_events and linear_get_issues to surface filings due in the next 14 days, then cross-references trading activity against documented risk limits. Replay covers every check, owner, and remediation step in the audit log.

P06

Run dashboard hygiene on a schedule

ReportingAnalyst validates row counts, null rates, and source freshness across every operational report, then attaches a data quality flag to each output. Scoped sql_query restricts reads to the schemas you list so hygiene checks cannot leak into production tables.

03
// The crew

Operations crew

Real personas from the operations crew. Each ships with a tuned system prompt and a default tool allowlist. Swap models per persona on the canvas.

Operations Manager

OperationsManager

Coordinates daily operations across the team, routes tasks to owners with deadlines, and surfaces blocked work before it slips.

Trading Operations

TradingOperations

Reconciles positions across CEX accounts, checks margin utilization, and flags settlement discrepancies with exact amounts.

Executive Assistant

ExecutiveAssistant

Owns calendars, drafts replies, preps meeting briefs from inbox context, and stages communications for leadership approval.

Reporting Analyst

ReportingAnalyst

Pulls live data from SQL and warehouse sources, validates row counts and freshness, and ships reports as Excel, slides, or chart files.

Data Operations

DataOperations

Watches API health and data feed freshness against named SLAs, then raises alerts the moment an endpoint or pipeline degrades.

Compliance Operations

ComplianceOperations

Tracks regulatory filing deadlines, maintains audit logs, and verifies trading activity stays inside documented risk limits.

Incident Manager

IncidentManager

Triages anomalies into SEV1, SEV2, or SEV3 tickets, escalates to the right channel, and drafts the post-mortem after resolution.

Operations Synthesis

OperationsSynth

Compiles every stream into a four-section executive brief sorted by criticality, then posts the final read-out to Slack.

04
// Scoped tools

Only the actions you grant.

Every tool below is a real shared tool from the Melaya bundle. Allowlist per agent; HITL-gate the writes; revoke any of them in one click.

shared/tools/melaya_exec/

Live position, balance, and fill data from the same exchange keys your trading desk uses. Reads are open to TradingOperations; melaya_place_order and melaya_cancel_order stay HITL-gated by default.

melaya_get_positionsmelaya_get_account_balancemelaya_get_open_ordersmelaya_get_ordersmelaya_get_my_trades
shared/tools/database/

Read warehouse and operational databases to assemble reports and freshness checks. Scoped DSNs are passed per agent so ReportingAnalyst only reaches the schemas you list; writes require an explicit HITL gate.

sql_querysql_schemasqlite_query
shared/tools/messaging/

Page on-call via Telegram or Discord when IncidentManager classifies a SEV1, with severity, owner, and resolution ETA. All outbound messages are HITL-gated for this crew by default.

telegram_send_messagediscord_send_messagediscord_send_embed
shared/tools/project_mgmt/

Open and update incident tickets, track filing deadlines, and route blocked work back to owners. jira_create_issue and linear_create_issue are HITL-gated so no ticket fires without an approver.

jira_get_issuesjira_create_issuejira_update_issuelinear_get_issueslinear_create_issue
shared/tools/calendar/

Pull the next 72 hours of meetings into the executive brief and stage holds for incident review calls. gcal_create_event is HITL-gated so nothing lands on a leader's calendar without sign-off.

gcal_list_eventsgcal_create_event
shared/tools/email/

ExecutiveAssistant reads inbound threads to prep meeting briefs and drafts replies for approval. gmail_send is HITL-gated by default; rules-based vendor responses can be lifted per template.

gmail_sendgmail_read
shared/tools/msoffice/

Ship reports as Excel workbooks, briefs as Word documents, and board read-outs as PowerPoint decks. File writes land in the run artifact log so any output can be replayed and traced.

excel_createexcel_write_datapptx_createword_create
shared/tools/core/

Vendor status-page polling via http_request, daily brief delivery via slack_post_text, and ad-hoc lookups via web_search. slack_post_text and file_write are HITL-gated by default for this crew.

web_searchhttp_requestslack_post_textfile_writegrep_search
05
// Three knowledge layers

The crew reads what you give it.

Every pipeline ships with three layers of knowledge access. Mix and match per agent on the canvas. No shared vector space with another tenant, no surprise reads, no opaque retrieval.

L1

Static context

includeContext

Per-pipeline documents appended to specific agents' input on every run. The ICP brief, playbook, pricing sheet, or won-deal email corpus. Whatever needs to be there before the agent thinks. You pick which personas get which docs.

L2

RAG retrieval tool

rag_retrieve

A scoped tool granted per-agent. When the agent decides it needs more depth, it queries the workflow's vector store on demand. Same knowledge base as Static context, accessed only when the model asks for it.

L3

Cross-run memory

pipeline_memory

Pipeline-level state that carries from one run to the next. Yesterday's research is in scope for today's follow-up. The crew remembers what it already prospected, what got approved, what was sent. The audit log is the second-order knowledge base.

07
// FAQ

Questions we get every week.

Will the operations crew act on its own?

No, not by default. Slack posts, Jira updates, Linear writes, and outgoing email are HITL-gated for this crew. The agent prepares the action and waits for one-click approval. You can lift the gate per template once you trust the run.

Can the crew reason over our runbooks and dashboards?

Three ways. Static context attaches your runbook, SLA matrix, and on-call rota to specific personas on every run. The rag_retrieve tool lets ReportingAnalyst and ComplianceOperations pull from data dictionaries and policy docs on demand. Cross-run memory carries yesterday's open incidents into today's stand-up so nothing resets at midnight.

Is this a Zapier alternative or a Make.com alternative?

It is closer to a Zapier alternative for the reasoning steps, and an n8n alternative for the schedule and webhook plumbing. Where Zapier and Make.com chain fixed actions, Melaya runs reasoning agents between the steps so an incident can be triaged, not just routed.

How do we keep reports from sounding like AI filler?

Every report cites the SQL query, the row count, the freshness timestamp, and the data quality flag. ReportingAnalyst is required to attach the source query and the staleness check before any output ships. The COO reads numbers and provenance, not adjectives.

Which models can we run the crew on?

Any. Claude on OperationsSynth where multi-source reasoning earns the cost, GPT on ExecutiveAssistant drafting, a local Ollama on TradingOperations when balance data must not leave your network. Each persona picks its own model.

How fast can ops get the first pipeline running?

With Slack, Jira, and a database DSN connected, the morning health-check pipeline is a 4-node canvas: trigger on cron, check API matrix, build status table, post to Slack. Most teams ship it in a working session and have the first brief in channel the next morning.

How does this handle on-call paging and incident severity?

IncidentManager classifies every anomaly as SEV1, SEV2, or SEV3 using the rules in your Static context. SEV1 routes to Telegram and Slack with the on-call owner tagged. SEV3 sits in the daily digest. The audit log captures the classification, escalation, and resolution timestamps.

Can I audit exactly what the agent did and why?

Every run logs every tool call, every model invocation, every approval decision, and every retry. Replay any run at any time. The audit trail covers the data source, the query, the freshness check, and the human approver, so compliance ops can reconstruct any morning brief.

Can we restrict which agents can write to Jira or Linear?

Yes. Tools are scoped per agent in the canvas. You can give DataOperations read-only http_request and sql_query, while only IncidentManager carries jira_create_issue and linear_create_issue. Scopes are enforced before the model even sees the tool.

How does it plug into our existing stack?

Native connectors for Slack, Telegram, Discord, Gmail, Google Calendar, Jira, Linear, Postgres, SQLite, Excel, Word, and PowerPoint. The melaya_exec bundle reads live positions and balances from the same exchange keys your trading desk already uses.

Build operations teams pipelines on Melaya.

Sandbox tier is free with no card. Join the waitlist and we will email you the moment a slot opens.

← Back to every use case
Join the community