mirrorsBook a demo
Backed byYCombinator

Tell your agent once.

Mirrors reminds your AI agent of your instructions at exactly the moment they matter, so it stops making the same mistakes

Book a demo
built for your agent’s context
  • LangChain
  • LangGraph
  • LlamaIndex
  • CrewAI
  • AutoGen
  • Pydantic AI
  • smolagents
  • Google ADK
  • Hermes
  • AI SDK
  • Mastra
  • OpenAI
  • Anthropic
  • MCP
  • Claude Code
  • Codex
  • Cursor
  • Windsurf

You can see the problem in your traces. Put your fix to work, without a redeploy.

Every session teaches you something. Where your agent gets stuck. What gets it back on track. Which context leads to good outcomes. Right now, those lessons stay buried in your traces.

Write the guidance you want your agent to follow. Mirrors routes it when your conditions match. Preview the sessions it will reach, compare the outcomes you track, and keep what works.

How it works.
From trace to fix in minutes, not sprints.

From experience to context at the right timeSessions, outcomes, and feedback reveal patterns in agent behavior. You write the guidance and preview where it belongs. Mirrors routes context when the situation matches, then compares outcomes with held-out sessions using the outcomes your team tracks.HOLDOUT / WITHOUT NEW CONTEXTGuidancecontextRulesroutingToolscontextTriggersscopeSessionsignalsResultscontextPromptguidanceReviewedlessons02 / CONTEXT THAT FITS01 / MIRRORS ENGINE03 / YOUR AGENT right context · right timeoutcome report04 / LEARN FROM THE OUTCOMEYOUR AGENT’S EXPERIENCEINPUT// a recorded session03 update_policy coverage: comprehensive// a customer correction“Check the license first.”// a successful resolutionverified · resolved · confirmedSessions · Outcomes · FeedbackEXAMPLEinformed byroutes contextcontext ↔outcomesholdout outcomes →Write the context.Use it next time.
01

Analyze your traces.

Review your agent’s sessions alongside your outcome labels and feedback. Decide where a context change could help.

02

Write the fix as context.

Author the guidance your agent needs and preview which sessions it would match, including unexpected ones, before going live.

03

Route it in real time.

When a matching situation comes up again, Mirrors delivers the context mid-conversation. No redeploy, no code change.

04

Measure the outcome.

Compare sessions with and without your context using the outcomes you track. Keep the changes that meet your goals.

See what happened. Fix it live. No code change required.

Every session has something to teach.

Review your traces and feedback to decide what context to change.

Know where your context will land.

Preview matching, missed, and unexpected sessions before activation.

The right context. Exactly when it matters.

Route guidance as the conversation unfolds, when its conditions match.

Give your agent a better next move.

Edit guidance, tool descriptions, schemas, and model-facing results.

Fewer repeat mistakes. Happier customers.

Compare context changes against the outcomes your team tracks.

What changes when your guidance reaches the right moment.
Before and after Mirrors.

TODAY WITH MIRRORS
01Successful sessions disappear into the logs.
↗What worked becomes evidence for the guidance you give next.
02Every failure adds another instruction to the global prompt.
↗Each lesson reaches the situations where it belongs.
03Your agent sees the same instructions as the conversation changes.
↗Relevant context arrives when the moment calls for it.
04Customers keep explaining the same mistake.
↗Measure which context changes reduce recurring failures.

711,933 agent sessions. And counting.

Turn past conversations into better agent performance. Find what works, measure the impact, and keep improving.

See what that looks like for your agent

Start with
one workflow.

Bring your traces and a workflow you want to improve. We’ll work through the context it needs and discuss pricing for your use case.

Let’s talk / pricing
01

Find the moments where better context could help.

02

Turn one recurring mistake into a lesson your agent uses.

Book a demo

FAQ

What is Mirrors?

Mirrors is a developer tool for AI agents. You turn what you see in your agent’s traces into guidance, and Mirrors routes that context into the agent mid-conversation when the situation matches. Before a change goes live, you preview which sessions it would reach. After, you compare outcomes with and without it. No redeploy or code change is needed.

How is Mirrors different from adding instructions to my system prompt?

A global prompt shows your agent the same instructions in every conversation, and every failure adds another line to it. With Mirrors, each lesson reaches only the situations where it belongs, at the moment it matters. Beyond guidance, you can edit tool descriptions, tighten input schemas, or rewrite tool results the model sees.

Does routing context slow down my agent?

Routing decisions run locally, without a network request or model inference to decide what applies. Configuration refresh and telemetry happen separately. If no valid configuration is available, your agent continues without new context edits.

Does Mirrors work with the traces I already have?

Yes. You start by reviewing your agent’s sessions alongside your own outcome labels and feedback. Routing conditions are previewed against stored traces, so you can check matches and misses before anything goes live, then compare sessions with and without your context against the outcomes your team tracks.

How much does Mirrors cost?

There is no public price list yet. Most teams start with one workflow: bring your traces and a change you want to try, and we scope pricing around that use case. Book a demo to talk it through.

Put your next context change to work. Right where it matters.

Bring a workflow, your traces, and a change you want to try. See where your guidance would apply and how to measure its impact.

Book a demo