Observability
Bonito vs LangSmith (LangChain)
Agent engineering platform vs. agent operations across providers
LangGraph Platform is now LangSmith Deployment, Agent Builder is now LangSmith Fleet, and the LLM Gateway is in public beta (July 2026)
LangSmith is LangChain's agent engineering platform: tracing and evals, LangSmith Deployment for running agents from any framework, Fleet for no-code agents, and an LLM Gateway in public beta. It is strongest for teams building on LangChain and LangGraph. Bonito puts a production gateway with shipped cross-provider failover and cost routing under the agents, on your own accounts across eight providers.
LangSmith pricing
Developer free (1 seat, 5K traces/mo); Plus $39 per seat per month; Enterprise custom. Deployment and Fleet usage billed in usage units.
Model costs
The gateway takes your own provider keys or Gateway Credits; any fee on credits is not published.
What LangSmith does well
- Tracing and evals tightly integrated with LangChain and LangGraph, and the deeper eval and prompt tooling of the two
- LangSmith Deployment runs agents from any framework, with cron jobs, human-in-the-loop, long-term memory, A2A and MCP
- Fleet builds no-code agents with Slack, Teams and Gmail channels and tool-level approvals in an agent inbox
- Compliance: SOC 2 Type II, HIPAA, GDPR and ISO 27001, with an EU regional instance
- Gateway spend limits at org, workspace, user and key level, with PII redaction and audit logging
Where Bonito goes further
- Automatic failover to an equivalent model on another provider, shipped; LangSmith lists gateway fallbacks as planned
- Cost routing to cheaper models for simple work
- Azure, Groq, OpenRouter and DeepInfra alongside Bedrock, Vertex, OpenAI and Anthropic
- Dollar budgets per agent, not only per org, workspace, user or key
- Managed knowledge bases with upload, embedding, search and a BYO bucket
- Data residency pinning to CA, US or EU that also constrains routing and failover
Where LangSmith is ahead of Bonito
- Eval depth and prompt management
- Native fit for LangChain and LangGraph code
- Compliance certifications (SOC 2 Type II, HIPAA, ISO 27001)
- Self-hosted LangSmith on Enterprise, and open-source LangChain and LangGraph
Capability by capability
| Capability | Bonito | LangSmith |
|---|---|---|
| Agents | ||
| Build and run agents | Yes | Yes |
| Approval queue and scheduled runs | Yes | Yes |
| Dollar budgets per agent | Yes | Partialgateway spend limits per org, workspace, user or key |
| MCP support | Yes | Yes |
| Routing | ||
| OpenAI-compatible gateway | Yes | Yespublic beta |
| Automatic failover across providers | Yes | Nofallbacks listed as planned, not shipped (Sep 2026) |
| Cheaper models for simple work (cost routing) | Yes | Nonot found in their docs (Sep 2026) |
| Bedrock, Vertex, Azure and direct providers in one place | Yes | PartialBedrock, Vertex, OpenAI, Anthropic and compatible endpoints; no Azure connector named |
| Semantic caching | No | Nonot found in their docs (Sep 2026) |
| Governance | ||
| SSO, RBAC and audit log | Yes | Yes |
| Data residency pinning | Yes | PartialEU regional instance for LangSmith data |
| Observability and knowledge | ||
| Agent traces with cost per request and agent | Yes | Yes |
| Knowledge bases (RAG) | Yes | Nonot found in their docs (Sep 2026); memory store, not a document KB |
| How you run it | ||
| Whole stack from one config file | Yes | Yeslanggraph.json and the LangGraph CLI |
| Self-hosted or open source | Partial | PartialLangChain and LangGraph are open source; LangSmith self-host is Enterprise |
| No markup on inference | Yes | YesBYO keys; fee on Gateway Credits not published |
Checked against LangSmith's own documentation on 2026-09-27. Products change quickly; if something here is out of date, tell us at hello@trybonito.com and we will fix it.
Sources
See Bonito on your own stack
Start free, or get a report on where agents fit in your company first.