Eventloom v1.0

A runtime trace layer for agent systems.

Preserve the human-to-agent session as a verifiable artifact. Eventloom records typed intentions, events, tool calls, model calls, approvals, and handoffs into an append-only JSONL log you can replay, inspect, and export.

Append-only JSONL SHA-256 hash chain Deterministic replay MCP + trace exports
Install locally Inspect a trace GitHub
runtime@eventloom/runtime mcp@eventloom/mcp reposyndicalt/eventloom releasev1.0 roadmap

Capture / Replay / Export

Agent work becomes typed runtime history.

Eventloom does not replace your editor, LLM provider, or observability stack. It sits beside them as the local record of what the agent proposed, what the orchestrator accepted, and what replay can prove later.

Capture

Write facts, not vibes.

Goals, task lifecycle, intentions, model calls, tool calls, approvals, and verification events land in append-only JSONL.

Replay

Rebuild state without rerunning.

Pure projections reconstruct tasks, actor mailboxes, effects, research state, event counts, and handoff context from history.

Export

Move traces where work happens.

Send the same local journal to MCP clients, Pathlight panels, HALO analysis, OTLP artifacts, CI bundles, or repo-adjacent logs.

Trace inspector

Paste JSONL, replay state, and generate handoff context.

Drop in an Eventloom log, paste the JSON output from eventloom visualize, or load the sample. The browser projects the same story the runtime sees: captured facts, replayed state, and the handoff a human or next agent needs.

Sample log loaded. Paste JSONL, upload a file, or paste visualize output.

sample.jsonl
Captured events Ordered facts from the log.

Session artifacts

The conversation that created the work belongs with the work.

Eventloom turns agent sessions into append-only records that can live beside code, CI artifacts, release notes, or handoff bundles. The transcript is useful; the typed runtime trace is what lets the next person or agent verify what changed.

Repo-adjacent journals

Keep local `.eventloom/*.jsonl` histories as development artifacts, fixtures, or archived release evidence.

Human-in-the-loop records

Represent approvals, rejections, and effects as explicit events instead of burying them in chat prose.

Agent handoff bundles

Rebuild active tasks, completed work, telemetry, observability gaps, and next actions from the same log.

Runtime boundary

Intentions are proposed. Events are accepted.

Actor intention
validate
Orchestrator
append
Hash-chained JSONL
replay
Projection + export

Install

Start with a local log file.

No hosted service, database, or daemon is required. Use the CLI directly or expose the same operations to agent clients through the MCP package and artifact bundles.

CLI

npm install @eventloom/runtime
npx eventloom append .eventloom/agent-work.jsonl \
  goal.created --actor user \
  --payload '{"title":"Ship the feature"}'
npx eventloom visualize .eventloom/agent-work.jsonl

MCP

npx @eventloom/mcp --root .

# tools: append, replay, timeline,
# explain task, mailbox, handoff,
# visualize, built-in workflows,
# Pathlight, HALO, OTLP

TypeScript

import { createRuntime } from "@eventloom/runtime";

const runtime = createRuntime(".eventloom/run.jsonl");
await runtime.runBuiltIn("software-work");
const visualizer = await runtime.visualize();

Observability

One local journal, multiple inspection surfaces.

Keep the canonical record local, then export when useful. Eventloom can turn agent journals into Pathlight traces with native Eventloom panels, HALO-compatible OpenTelemetry JSONL, generic OpenTelemetry OTLP JSON, or vendor-neutral OTLP trace JSON.

Pathlight / HALO / OTLP

npm run eventloom -- run software-work /tmp/eventloom-pathlight-viz.jsonl
npm run eventloom -- export pathlight /tmp/eventloom-pathlight-viz.jsonl \
  --base-url http://localhost:4100 \
  --trace-name eventloom-pathlight-viz
npm run eventloom -- export otlp /tmp/eventloom-pathlight-viz.jsonl \
  --out /tmp/eventloom-otlp-traces.json
Pathlight guide HALO guide OTLP guide User guide