LangSmith engineers your agent. We measure your product and your agent.
Different jobs, and we will say so plainly: LangSmith is deeper than us on agent tracing, evals and deployment, it is the native choice for a LangChain stack, and it has $125M of Series B behind it. smolanalytics is one MIT Go binary that does product analytics and computed agent observability from the same events, self-hostable by anyone, with no metered AI.
smolanalytics vs LangSmith, which one do I want?
It depends on what you are building, and plenty of teams run both. LangSmith is LangChain's commercial platform for agent engineering: tracing and monitoring of every agent run, evaluation, a playground, and deployment of agents to production, all native to LangChain and LangGraph. It is deeper than smolanalytics at all of that, it is the default choice if your stack is built on LangChain, and the company raised a $125M Series B at a $1.25B valuation in October 2025, the same month it shipped LangChain 1.0 and LangGraph 1.0. Two facts to weigh before you pick it: the platform itself is closed source (the SDKs and the LangChain framework are MIT), and self-hosting is an add-on to the Enterprise plan, in their own words designed for their largest, most security-conscious customers, running on ClickHouse, PostgreSQL and Redis. Cloud pricing is per seat plus per trace: a free Developer seat with 5k base traces a month, Plus at $39 per seat per month with 10k base traces, then pay-as-you-go, with base traces retained for 14 days. smolanalytics is a different shape: a single MIT Go binary that does web and product analytics (funnels, retention, paths, cohorts, breakdowns, heatmaps, surveys, feature flags and A/B) and computed agent observability off the same event stream, tool-call volume, error rate, latency p50/p90/p99, client split, an error taxonomy, and conversation health like turns, re-ask rate, abandonment and resolution. Those numbers are computed from your events, never labelled by a model. You ask it questions in plain English from your editor: one MCP connection with an org token operates every project, 79 tools, answered by your own coding agent on your own model, so the AI is never metered by us. Pricing is a 14-day full trial with no card, Pro $19/mo for 2M events, Scale $49/mo for 5M, $6 per million, self-host free forever.
what each one actually is
“Know what your agents are really doing”: an Agent & LLM Observability Platform with tracing, monitoring and insights, plus evaluation and deployment, from the LangChain team.
langchain.com/langsmith
“Self-hosted LangSmith is an add-on to the Enterprise plan designed for our largest, most security-conscious customers.” It runs on ClickHouse, PostgreSQL and Redis, plus its own frontend, backend, platform backend, playground and queue services.
docs.langchain.com/langsmith/self-hosted
smolanalytics is one MIT Go binary and one data file that anyone can self-host, on any plan or none. Product analytics and agent observability come out of the same events: no seats, no trace meter, no cluster.
what we do instead
LangSmith vs smolanalytics
| LangSmith | smolanalytics | |
|---|---|---|
| The job it is built for | agent traces, evals, deployment | product analytics + agent health |
| Trace-level agent debugging | deep: tracing, monitoring, insights | no, we compute rollups |
| Evals and datasets | core product | not offered |
| Prompt playground | yes | not offered |
| Deploying agents to production | yes, that is half the platform | not offered, we only measure |
| Funnels, retention, paths, cohorts | not its category | yes, same binary, same events |
| Pricing model | per seat plus per trace | per event, no seats |
| Self-host | Enterprise add-on: ClickHouse + PG + Redis | one static binary, one file, free |
| Dashboard vs API drift | separate read paths | CI pins /v1 == MCP == dashboard |
| License | platform closed source, SDKs MIT | MIT, all of it |
when to use LangSmith (often alongside us)
- You need to debug a single agent run step by step. LangSmith's whole observability product is tracing, monitoring and insights over agent behaviour, in their words “see exactly what your agents are doing”. We compute rollups over your events and do not store per-run traces, so if the question is “what happened inside that one run”, that is LangSmith, not us.
- Your stack is LangChain or LangGraph. LangSmith is built by the same team, is native to both frameworks, and covers evals, a playground and one-click deployment on top of tracing. We ship none of that and are not trying to. Running both is a normal setup: LangSmith for the agent's lifecycle, smolanalytics for the product and the computed agent health beside it.
- You are an enterprise that wants a vendor. Custom SSO and RBAC, self-hosted and hybrid deployment options, and a company that raised $125M at a $1.25B valuation in October 2025 to build exactly this. That is a real reason to pick them, and self-hosting being gated to that Enterprise plan may not matter to you at that size.
Add the other half in 30 seconds
# no ClickHouse, no Postgres, no Redis, no Enterprise plan to self-host docker run -p 8080:8080 -v $PWD/data:/data \ -e SMOLANALYTICS_DB=/data/smolanalytics.data \ ghcr.io/arjun0606/smolanalytics # dashboard + MCP at http://localhost:8080 # then, in your editor: "re-ask rate and p90 latency for the agent this week, by client"
Keep LangSmith for traces, evals and deploys if you build on LangChain. Put this next to it for funnels, retention and the computed agent numbers, on your own box, answered by your own model.
questions
Is smolanalytics a LangSmith alternative?
Only partly, and we would rather be straight about it. LangSmith is LangChain's platform for agent engineering: tracing, monitoring and insights over agent runs, evals, a playground, and deployment, and it is deeper than us at every one of those. smolanalytics is one MIT Go binary that gives you web and product analytics (funnels, retention, paths, cohorts, heatmaps, surveys, flags and A/B) plus computed agent observability from the same events: tool-call volume, error rate, latency p50/p90/p99, client split, an error taxonomy, and conversation health like turns, re-ask rate, abandonment and resolution. If trace-level debugging, evals or deployment are what you need, use LangSmith. If you want product usage and agent health in one place you can self-host with no gates, use us. Many teams run both.
Can smolanalytics replace LangSmith for tracing, evals or deployment?
No. We do not store per-run agent traces, we do not ship evals or a prompt playground, and we do not deploy your agents, and we are not planning to. Our agent observability is computed rollups over the events your app already sends, which answers “is the agent healthy and where is it failing”, not “show me every step of run 4f2a”. For that second question, LangSmith.
Is LangSmith open source, and can I self-host it for free?
The LangSmith platform is closed source; the client SDKs and the LangChain and LangGraph frameworks are MIT. Self-hosting LangSmith is, per their own docs, an add-on to the Enterprise plan designed for their largest, most security-conscious customers, and the deployment runs on ClickHouse as the primary trace store plus PostgreSQL and Redis. smolanalytics is MIT all the way down, and self-hosting it is one Go binary and one data file, free forever on any plan or none.
How are smolanalytics agent numbers different from LangSmith's?
Ours are computed, not labelled. Tool-call counts, error rate, latency percentiles, client split, the error taxonomy, turns, re-ask rate, abandonment and resolution are all derived deterministically from your event stream, so nothing is a model's opinion and nothing changes when a model changes. LangSmith works over per-run traces and also ships an Insights product that surfaces usage patterns and issues automatically; if you want that trace-level depth, that is their strength, not ours.
What does smolanalytics cost, and can I self-host it?
There is a free tier (1M events/mo, no card) and a 14-day full trial of Pro on top of it. Pro is $19/mo for 2M events, Scale is $49/mo for 5M, and extra usage is $6 per million. Self-hosting is free forever under MIT: one Go binary, one data file, no external services. There is no free cloud tier.
Get the verdict, not another dashboard.
Drop one snippet, ask in your editor, own your data. 14-day full trial, no credit card.