Battle-tested at Spotify. Yours to try.

Know your systems. So your agents can too.

Xirp connects to your services, ownership, docs, and architectural decisions — so every AI coding session starts with real context, not guesswork.

Join the beta Read the docs

The product

One environment.
Two surfaces, same memory.

Run sessions in the Xirp desktop app. Keep the context in Portal — your own instance is included with the beta.

The Xirp desktop app: a session list beside an active Claude Code session with project context

01 — output

AI coding tools solved the generation problem.

Code ships faster, more services get built, output goes up.

02 — but

Agents make fast, confident decisions that are technically correct — and operationally wrong.

Because they don't know what they're actually working inside.

03 — the knowledge exists

It's in Slack threads nobody could find. In the heads of the three people who built that service in 2021.

The knowledge that would fix this was never missing.

04 — the real diagnosis

That's not a documentation problem. It's a retrieval problem.

READMEs, Confluence pages, and architecture diagrams just go stale while the real work keeps moving.

05 — xirp

Xirp is different. Your agents don't guess. They know.

An agentic development environment that understands your code — connected to Portal, it knows your services, ownership, dependencies, and architectural decisions. All of it, in every session.

The difference

Same prompt. Same model.
A different answer.

Flip the switch to see what changes when the agent actually knows your system.

prompt: “Add retry logic to the checkout call.”

agent — no context

✓ wrote exponentialBackoff(retries: 3, base: 250ms) around checkout.charge()

✓ added jitter, capped at 4s — textbook implementation

checkout already retries at the gateway — double retries caused the March incident
violates team-spectra's error-handling standard (ADR-0098)
nobody will notice until the next traffic spike

Technically correct. Operationally wrong.

Four pillars

Institutional memory,
engineered four ways.

Agent

The agentic development environment with institutional memory.

Other agents see the file you're working in. Xirp sees the system it's part of: who owns the upstream service, what depends on it, why it was built this way. Every session starts grounded, and every session makes the next one smarter.

owner who owns the upstream service
dependents what breaks if you change it
decisions why it was built this way

A Xirp session grounded with project overview, tech stack, and entry points

Workspace

The context layer your agents and engineers share.

Workspace, a Portal plugin, holds your work items, sessions, and docs in one place. When an engineer wraps a session, the context doesn't disappear. It's stored and ready for the next person, or agent, who touches that system.

Workspace in Portal: a sessions list showing recent work by engineers and agents

Documentation

Living documentation, built automatically.

Every coding session generates real knowledge. Xirp captures it, turns it into documentation that stays current, and feeds it back into future sessions so every agent and engineer starts smarter. No more asking engineers to stop and write docs. It happens as the work happens.

session→knowledge→docs→smarter next session↺

docs/checkout-retries.md — auto-generated

A living documentation index generated from real coding sessions

Flexibility

Works with Claude, Gemini, Codex. Knows your org, not theirs.

Xirp is a harness for the model you prefer. Switch between Claude, Gemini and Codex without losing context. Run sessions locally or remotely. Stay vendor-independent while the context layer compounds with every session.

modelClaudeswapped — nothing lost

runs /

context layer87%

your org's memory — it stays, and compounds with every session

Xirp running with different models while the same project context persists

Battle-tested
at Spotify

Built and run inside Spotify, against real systems at real scale.

Your own
Portal instance

Beta access includes Spotify Portal for your team — the context layer's home.

Model-
agnostic

Claude, Gemini, Codex. The context layer is yours, whoever writes the code.

Get started

Start with one team.
One project.

See what your agents can do when they actually know your systems.

Join the beta Join Portal Office Hours live demos every week

Watch the webinar: Powering your agentic transformation