LangSmith

https://www.langchain.com/langsmith

Memory reads, vector DB retrievals, state changes are distinct span types in traces. RAG eval separates retrieval quality (context precision) from generation quality (faithfulness). Dataset versioning guards against eval drift.

At a glance

Type
Memory mutations as distinct span types
Tier
T1
Created
2023 (LangSmith beta summer 2023; GA Feb 2024 with $25M Series A)
Latest release
not applicable — not OSS
License
not applicable — not OSS
GitHub
not applicable — no GitHub repo
Pricing
Developer: free (5K traces/mo); Plus: $39/seat/mo (10K traces); Enterprise: custom (SSO/SAML; dedicated support)
Funding
LangChain total ~$35M+ ($3M seed Benchmark 2023; $25M Series A Sequoia Feb 2024; $125M Oct 2025 unicorn round at $1…

Taxonomy

storage
relational
retrieval
exact-match
persistence
long-term
update
append-only
unit
episode
governance
auditable
conflict
n/a

When to use

Optimised for: memory operation tracing + drift / poisoning detection

Anti-fit: not for use cases that don't run agent workloads in production

Pros & cons

Pros

Best integration with the LangChain / LangGraph stack — debug memory + chain + agent in one trace; the de-facto trace tool for that ecosystem.

Cons

LangChain-shaped — works less well outside that ecosystem; closed SaaS.

Claims & capabilities

Free 5k traces/month Developer; Plus $39/mo; Enterprise custom. No automated drift / poisoning detection.

Technical surface

API surface
REST, SDK: Python, JS/TS
Backend storage
custom (LangChain-managed)
Deployment
SaaS cloud (US); Enterprise: dedicated tenant
Embedding model
not applicable — observability product
Multi-tenancy
namespace (workspace/project)
MCP
via LangChain MCP adapters
A2A
no data — searched langchain.com/langsmith, docs.smith.langchain.com, docs.smith.langchain.com/observability/concepts; A2A protocol not advertised (OTel-based traces, webhook automations)
OpenTelemetry
first-class — OTel ingestion + export

Compare LangSmith with…

Similar systems

Other memory observability & monitoring in the catalog, ranked by inbound references.

  • AgentOps T2

    When Mem0 is connected, gains Memory Operation Timeline, Search Analytics, Memory Growth tracking, Error Tracking per memory call. Standalone, records context at each step but doesn't analyse memory quality.

  • Galileo (galileo.ai) T1

    Treats memory as first-class in multi-agent tracing. Luna-2 SLMs (3B / 8B) scan every interaction for intent drift and belief drift; 20+ checks at sub-200ms latency. Catches when agent A's view of the world splits from teammate B's. OpenTelemetry-compatible.

  • Langfuse T1

    Memory module reads/writes captured as named spans. Trace Log View concatenates every agent step including memory ops. Agent Graphs (GA 2025) infer graph structure from observation nesting; session-level replay tracks how memory state evolves.

  • Ratine T2

    Only tool found that scans the persistent memory layer on disk rather than runtime tracing. Detects injected instructions, obfuscated payloads (zero-width Unicode, base64, homoglyphs, hex), C2-pattern URLs, credential leakage. ratine diff compares snapshots for belief drift.

Related systems

References (3)

  • LangChain (framework) integrates with — Best integration with the LangChain / LangGraph stack — debug memory + chain + agent in one trace
  • LangChain (framework) depends on at runtime — backend-storage cell: custom (LangChain-managed)
  • LangGraph Persistence integrates with — de-facto trace tool for that ecosystem

Row last verified 2026-05-14. Catalog data is CC-BY-4.0 — see how to read this.