Ingest raw evidence
bitter log put --bucket agent_runs --run-id 5468 --agent builder tmp/agent-runs/builder_5468.jsonlPreserve the original trace first. Parse later.
curl https://api.bitterlog.com/api/v1/status
BitterLog is the append-only evidence workbench for agent-run proof. It stores raw traces, receipts, prompt/output artifacts, and annotations with stable digests so operators can move from a run summary back to the exact source record.
Account-scoped access runs through Bitter. ClickHouse-backed search is a projected read model, not the ledger itself, and the public API status tells operators when analytics is available.
API and CLI ingest
Post operating events, upload raw traces, or bring the receipts your agent stack already emits.
Receipts stay intact
Prompts, stdout, stderr, manifests, and annotations keep stable digests and timestamps you can inspect later.
Account-scoped beta
One Bitter account scopes evidence, receipts, and annotations without introducing a BitterLog-only login.
evidence ledger
01 / CLI
BitterLog is optimized for the way agent systems actually behave: lots of raw logs, repeated research, and constant need to pivot from a summary back to the original evidence.
Ingest raw evidence
bitter log put --bucket agent_runs --run-id 5468 --agent builder tmp/agent-runs/builder_5468.jsonlPreserve the original trace first. Parse later.
Inspect and pivot
bitter log list --bucket agent_runs --run-id 5468
bitter log get --bucket agent_runs --id 71 --body-onlyMove from run identity to the exact bytes that explain what happened.
Check projected reads
curl https://api.bitterlog.com/api/v1/status
curl https://api.bitterlog.com/api/v1The public API says which read paths are available. Analytics are a projected model, not the ledger.
02 / Identity
Session JSONL, stdout, stderr, prompt snapshots, rendered output, and receipts stay intact with stable digests and timestamps.
Use the public API index and status to see available read paths. Projected analytics stay separate from the source ledger.
Attach annotations, preserve judgment, and turn recurring agent behavior into evidence you can act on.
03 / Trust
BitterLog keeps source evidence separate from downstream analysis. Raw evidence remains the durable ledger while projected query paths, when enabled, explain what happened without rewriting the source record.
Source record stays inspectable
Raw traces stay tied to stable digests and timestamps. Projected reads work from refs and projections, not a rewritten source of truth.
Reviewable receipts
Each run can connect request context, stdout, stderr, annotations, and deploy receipts so operators can audit the same path.
Read-only query path
When analytics is enabled, projected query surfaces explain patterns over evidence without mutating the append-only ledger.
04 / Research
Use the public API index and status to see available evidence paths. ClickHouse sits behind BitterLog as a projected read model for structured analysis when analytics is enabled; it is not the append-only source ledger.
starter moves
bitter log tables
05 / Access
Bitter is the account authority. BitterLog remains headless-first: the CLI or a constellation service exchanges an account assertion for a short-lived credential bundle, then writes receipts under your account scope.
launch access
Start from the Bitter account bridge, then use the public API index and status route to confirm which evidence paths are available. Evidence onboarding remains request-reviewed during launch.
Check API statusPrefer email? founders@bitterlog.com