v1.0 — Production-ready enterprise agent runtime

The enterprise AI agent platform, ready on day one.

AIOps Bot Platform ships a complete, auditable agent loop — tools, skills, a workspace file system, context compaction, and full observability. Deploy agents anywhere, keep secrets locked down, and bring your own LLM to go limit-free.

Get started freeRead the docs
AES-256 + VaultSOC2-alignedOIDC SSOAudit-grade logging
aiops-bot-platform — agent loop · prod-log-watcherlive
agent.perceive("prod-log-stream")
> 3 anomalies detected · queue growing on payments-svc
agent.reason("classify + plan remediation")
> plan: read trace > patch config > open PR
agent.act([fs.write, git.pr, slack.notify])
PR #482 opened · autoscaler config bumped
compacted 14k tokens · 0 secrets leaked · 412ms p50
observing… next cycle in 2s

Fits into your existing enterprise stack

HashiCorp VaultOIDC / SSOPostgreSQLKubernetesOpenTelemetryDatadogSlackGitHubAWSGCPAzurebrowser-useHashiCorp VaultOIDC / SSOPostgreSQLKubernetesOpenTelemetryDatadogSlackGitHubAWSGCPAzurebrowser-use
The platform

Everything an agent needs. Nothing you have to build.

AIOps Bot Platform is the full runtime, control plane, and observability layer — so your team ships agents, not infrastructure.

Ready-made agent loop

Perceive, reason, act, and observe — out of the box. Tools, a workspace file system, context compaction, and structured tracing are wired in, not bolted on.

  • Tool calling
  • Context compaction
  • Tracing

Deploy agents anywhere

Run a fleet on a server triaging production logs, or a single assistant on a user's laptop. Multiple agents can follow one project and stay in sync.

  • Server / edge / device
  • Multi-agent projects

Worry-free security

Secrets are encrypted in transit and at rest with AES-256-GCM. Plug in HashiCorp Vault for centralized key management and rotation.

  • AES-256-GCM
  • Vault integration
  • Per-agent scoping

Global skills repository

A shared, versioned catalog of skills your agents can pull on demand. Compose, reuse, and govern capabilities across every team.

  • Versioned catalog
  • Org-scoped governance

Browser-use skills

Give agents reliable web automation with first-class browser-use skills — scraping, form flows, and UI validation, sandboxed by default.

  • Headless automation
  • Sandboxed runtime

Bring your own LLM

Route to any provider or self-hosted model. Supply your own keys to go limit-free, with per-conversation overrides and cost guardrails.

  • Any provider
  • Limit-free
  • Per-conv overrides
Explore every capability in the docs
The agent loop

A complete loop, already wired.

The hard part of production agents isn’t the model — it’s everything around it. AIOps Bot Platform handles the loop so you focus on the behavior.

01

Perceive

Ingest logs, events, files, or live web pages.

02

Reason

Plan multi-step work with tool-aware memory.

03

Act

Execute tools, write files, open PRs, call APIs.

04

Observe

Verify results, capture traces, re-plan if needed.

Loops continuously, compacts automatically, observes its own output

Tools, on tap

Typed, schema-driven tools for the filesystem, shell, git, HTTP, and your internal APIs — with approvals and allow-lists.

Workspace file system

Every project gets a sandboxed, audited workspace. Agents read, write, and version files without touching host storage.

Context compaction

Long-running agents stay coherent. AIOps Bot Platform compacts and summarizes context automatically to fit any model window.

Observability built in

Token usage, tool calls, latency, and errors are traced per request — exportable to OpenTelemetry, Datadog, or your stack.

Run it anywhere

One runtime. Anywhere an agent is useful.

The same skills, tools, and security model follow your agents — whether they live in your cluster or on someone’s laptop.

Server

On the server, watching production

Deploy a fleet that tails logs, correlates incidents, and opens remediation PRs — 24/7, with scoped tokens and full audit trails.

Device

On the user's device, assisting daily

Ship a local assistant that reads the user's files and context while keeping data on-device. Same loop, same skills, narrower blast radius.

Fleet

Many agents, one project

Attach multiple agents to a single project so specialists collaborate — a triager, a fixer, and a reviewer sharing one workspace and memory.

Security first

Worry-free security,
by default.

Agents that touch production need production-grade secrets handling. AIOps Bot Platform treats every credential as compromised-by-default and locks it down accordingly.

secret-store/v1/payments-svc
STRIPE_KEYsk_live_••••••••••••e4a9
DB_URLpostgres://••••••@db
rotationvault · every 24h
Decryptable only inside the agent sandbox · zero exposure to logs

Encrypted in transit & at rest

Every secret is encrypted with AES-256-GCM using keys derived via scrypt. Tokens are hashed, never stored in plaintext.

HashiCorp Vault integration

Bring Vault as your secrets backend for centralized management, dynamic secrets, and automatic rotation — no code changes.

Per-agent secret scoping

Each agent gets a least-privilege token. Revoke, rotate, or scope down in a click — all actions audited.

Audit-grade logging

Immutable records of who did what, when. Ship to your SIEM for compliance with SOC2, ISO 27001, and internal policy.

Skills & browser use

A shared skills repo, browser included.

Skills are versioned, composable capabilities your agents pull on demand. Use the global repository, fork your own, and add first-class browser-use skills for reliable web automation.

  • Versioned, org-scoped governance for every skill
  • First-class browser-use: scraping, form flows, UI validation
  • Sandboxed runtimes with per-skill resource limits
Browse skills registry
Global skills repository128 skills
git.pull-requestfs.searchbrowser.form-flowsql.queryhttp.fetchbrowser.scrapelog.tailslack.notifypdf.extractsentry.lookupkubectl.diffemail.draft+ your own
browser-use skill
sandboxed

agent.use("browser.scrape", { url }) → structured DOM

Bring your own LLM

Any model. Zero limits.

Route to OpenAI, Anthropic, Google, or your own vLLM deployment. Bring your own keys to bypass platform quotas entirely — with per-conversation overrides and cost guardrails that keep you in control.

OpenAIAnthropicGooglevLLMLiteLLMOllama
aiops-bot-platform.config.ts
model: "bring-your-own",
provider: "anthropic",
apiKey: vault("ANTHROPIC_KEY"),
overrides: {
  conv-482: { model: "qwen-qwq-32b-high" }
}, 
limits: null // your key, your rules
99.95%
Control plane uptime
412ms
Median tool round-trip
0
Secrets exposed to logs
∞
Requests on your own key
From zero to agent

Live in an afternoon, not a quarter.

No agents to bootstrap, no infra to glue together. Create a project, attach an agent, and watch it work.

  1. 01

    Create a project

    Spin up a workspace, invite teammates, and scope secrets.

  2. 02

    Attach an agent

    Generate a scoped token and point your agent at the runtime.

  3. 03

    Ship to production

    Observe, iterate, and scale — all from one console.

Follow the quickstart guide
terminal — spin up an agent
$ npm create aiops-bot-platform@latest my-agent
$ cd my-agent&& aiops-bot-platform link
✓ project "payments-triage" linked
✓ agent token scoped · vault wired
$ aiops-bot-platform deploy
✓ live · 2 agents watching prod-logs
streaming telemetry…

Ship your first production agent today.

AIOps Bot Platform is the all-inclusive runtime, control plane, and observability layer for enterprise AI agents.

Get started freeBrowse the docs

No credit card · Enterprise SSO · Self-host or managed