What does Datadog LLM Observability cost?
There is a free tier of 40,000 spans, with Pro from $160/month, priced per span and with an LLM activation premium of around $120/day.
Beam vs DATADOG
Beam vs LLM Observability
Datadog LLM Observability puts agent and LLM tracing inside the monitoring platform many teams already run, with a 40,000-span free tier and Pro from $160 a month plus per-span costs and an LLM activation premium of about $120 a day. It is production-application observability. Beam is a local security monitor for coding agents on developer machines.
Beam vs LLM Observability, side by side
Details below are drawn from a September 2026 review of the AI agent security market. Some figures are vendor-sourced or analyst estimates, and this is a fast-moving space — check with each vendor before you buy. Composio details come from its own pricing and documentation pages as of October 2026.
| Beam | LLM Observability (Datadog) | |
|---|---|---|
| What it is | Local-first activity monitor, policy enforcement, and pre-run scanner for AI coding agents, with a Sentinel browser extension for AI chats. | LLM and agent observability inside Datadog's monitoring platform. |
| Monitoring | Shell commands, file changes, tool calls, network endpoints and browser actions from harnesses such as Claude Code, Codex and Cursor, plus Beam Sentinel checks on supported AI chat sites before send. Normalised into local NDJSON with risk flags and evidence attached. | Traces, agent steps and token spend for production AI applications, inside Datadog. |
| Skill / MCP scanning | Yes. SKILL.md files and MCP configurations are scanned before you run them, using 11 heuristic patterns plus an MCP version-pin check. Workspace admins can also block an MCP server for the org or one user. | Not applicable — different category. |
| Blocking / enforcement | Yes. Workspace owners and admins set observe, advisory, or enforce policy. Enforce denies matching tool calls on Claude Code and Codex PreToolUse hooks. The Sentinel Chrome extension pauses or blocks high-risk browser sends and can hold a send after redacting live secrets. | Not applicable — observability rather than endpoint enforcement. |
| Individual use | Free for up to 3 people, and Pro for one person at $8/month. Team workspaces add owner/admin roles, org baselines, stricter member overrides and a control plane. | Free tier of 40,000 spans. |
| Deployment | Local collector plus a workspace dashboard. Apple and Windows MDM. Sentinel Chrome extension for browser AI chats. Licensed self-hosted control plane for teams that need it on their own infrastructure. | Hosted, within an existing Datadog deployment. |
| Source / licence | AGPL-3.0 for the local CLI and collector; the multi-user control plane is proprietary. | Commercial. |
| Pricing | Free for up to 3 people. Pro $8/month for one person. Team $20 per developer and $5 per browser seat a month, billed yearly. | Free tier of 40,000 spans; Pro from $160/month, priced per span with an LLM activation premium of about $120/day. |
LLM Observability in numbers
As reported in our source research. Some figures are estimates.
- Free tier
- 40,000 spans
- Pro
- From $160/month
- Additional
- Per-span pricing plus an LLM activation premium of about $120/day
Where LLM Observability does more
- Lands inside a monitoring platform many organisations already run, with no new vendor to approve.
- Mature tracing across services, not just the AI layer.
- A free tier of 40,000 spans to start with, and published pricing.
Where Beam differs
- Beam is about a developer's machine and the commands an agent ran there, with risk flags and evidence.
- Beam scans skill files and MCP configurations before you run them.
- Beam is local-first with no per-span bill and no data leaving the machine by default.
Beam’s current gaps
- Enforce is hook-based: a native deny is wired for Claude Code and Codex PreToolUse. Cursor, Copilot CLI and Gemini currently get a deny note, not a verified block.
- Missing, unreadable or expired cached policy allows the action. Beam is not a fail-closed kernel sandbox or an in-silicon watchdog.
- Scanning is heuristic pattern matching. It produces false positives, misses novel phrasing, and is not a guarantee of safety.
- There is no SOC 2 report and no default SIEM forwarder. Kernel isolation like OpenShell remains a different layer.
Which one fits
If your production AI already runs through Datadog, adding LLM Observability there is the obvious move. Beam does not compete with it — it answers a security question about the endpoint, with workspace policy and Sentinel, and it has no SOC 2 report.
FAQ
Does it monitor local coding agents?
It targets production AI applications rather than local coding-agent safety, which is the surface Beam covers.
Can Beam export to a monitoring platform?
Not today. Beam has no SIEM or platform forwarding — it writes redacted NDJSON locally and offers a local export for investigation.
See how Beam works
Beam is an early, local-first prototype. Read what it actually does today before comparing it further.