# Aient

> Aient is an AI SRE (site reliability engineer). It detects production problems from your OpenTelemetry data, root-causes crash-class bugs against your actual source code, and opens a reviewed GitHub pull request with the fix — then, after you merge, verifies in production that the problem is resolved and reopens it if it regresses.

Aient is built for engineering-led teams that ship to production daily but have no dedicated SRE org. It is OpenTelemetry-native (CNCF-standard, no proprietary lock-in) and human-in-the-loop by default: Aient opens the pull request, and you keep the merge button.

Operated by Triple Alpha AB. The product is in early access (alpha).

## What Aient does

- Detects problems broadly: crashes, error spikes, and latency or availability anomalies, from OpenTelemetry traces and logs.
- Groups related signals into logical issues instead of raw alert noise.
- Root-causes the code-level subset — exceptions and crashes that resolve to a stack frame in a repository it can read — against your actual source code.
- Opens a reviewed GitHub pull request with the proposed fix. You review every fix before it ships.
- After you merge, verifies in production that the error actually stopped, over a short configurable window, and reopens the problem if it regresses.

Infrastructure, configuration, capacity, data, and upstream-vendor causes are detected and routed to your team with full context rather than auto-fixed.

## How it is priced

A flat $99/month base plus usage for the AI work — no per-seat fees, no per-gigabyte telemetry tax, and no per-repo limits. You set a budget and Aient never exceeds it, with live spend tracking and notifications at your threshold. There is no free tier; a card-gated warm-up trial is the way in.

## Data and privacy

Your telemetry is stored in EU infrastructure (EU-region Supabase and object storage). Only the remediation step uses an external sandbox and large language models (OpenAI / Anthropic), under standard contractual clauses. Aient never sells your data and never shares it with model providers for their training. A SOC 2 report is available on the Enterprise tier (SOC 2 is in progress). Each customer runs in a single-tenant sandbox with organisation-scoped data isolation.

## MCP server (Model Context Protocol)

Aient is MCP-native: the same AI SRE is reachable over the Model Context Protocol, so agents in Claude, Cursor, or any MCP client can ask what's broken in production and act on it. The server is remote and HTTP-based at https://aient.ai/mcp (a wallet-paid transport for agentic onboarding is at https://aient.ai/mcp-x402); nothing to install or self-host. It exposes more than 30 tools — including list_problems, get_problem, query_traces, query_logs, get_spans, detect_anomalies, service_health_summary, and request_problem_remediation. Access is scoped: read-only by default (aient.mcp.read scope), and mutations such as requesting a remediation require aient.mcp.write. A remediation opens a reviewed GitHub pull request — it merges only after explicit human approval on the exact PR head (merge automation is opt-in per environment, off by default). The published server manifest is at https://aient.ai/.well-known/mcp/server.json (registry identity ai.aient/mcp).

## Key pages

- [Aient home](https://aient.ai/): What Aient is and how the closed loop works.
- [Pricing](https://aient.ai/pricing): $99/month base plus usage, budget controls, and an FAQ.
- [MCP server](https://aient.ai/integrations/mcp-server): Connect Aient over the Model Context Protocol from Claude, Cursor, or your own agent — 30+ tools, read-only by default, remediation opens a reviewed PR.
- [ROI calculator](https://aient.ai/roi-calculator): Free tool — estimate the engineering hours and cost an AI SRE gives back, from your own incident volume, hours per incident, and code-level share.
- [Compare](https://aient.ai/compare): How Aient compares to other AI SRE and autonomous remediation tools.
- [Aient vs Sentry Seer](https://aient.ai/compare/aient-vs-sentry-seer): Standalone and OpenTelemetry-native vs a per-active-contributor add-on inside Sentry.
- [Aient vs Resolve.ai](https://aient.ai/compare/aient-vs-resolve-ai): Transparent, self-serve pricing vs enterprise sales-only.
- [Blog](https://aient.ai/blog): Engineering deep-dives, postmortems, and announcements; RSS at https://aient.ai/blog/rss.xml. New posts are added here as they publish.
- [Privacy policy](https://aient.ai/legal/privacy): How Aient handles your data.