> ## Documentation Index > Fetch the complete documentation index at: https://opensre.com/docs/llms.txt > Use this file to discover all available pages before exploring further. # Session History > Browse recent sessions, resume a conversation, compact long context, or start fresh. ## Commands at a glance | Command | What it does | | - | - | | `/sessions` | Browse recent sessions and resume the highlighted one | | `/resume ` | Restore a past session so you can continue asking questions | | `/resume :` | Restore a specific branch point by entry ID prefix | | `/compact` | Summarize older context into a replayable compaction entry | | `/context` | Show what the next model call will contain and how much of the history budget is used | | `/new` | Start a fresh session file while keeping the current conversation context | *** ## Browsing past sessions — `/sessions` ``` /sessions ``` In a terminal, **↑/↓** moves the highlight, **Enter** resumes that session, and **Esc** closes the picker. It shows up to 20 sessions with relative times. The status line shows the selected session's ID, start time, duration, and turn count. The current session is marked, and the most recent previous session is highlighted by default when one exists. When output is piped, `/sessions` prints a table with IDs, start times, durations, and turn counts. Use the displayed ID with `/resume `. The current session updates its elapsed time live. *** ## Resuming a past session — `/resume` Use `/resume` with the first characters of the session ID from `/sessions`: ``` /resume 9b2e4f7a ``` From outside the REPL, pass the same ID or name substring to the CLI launcher (for example `opensre --resume 9b2e4f7a`, or `o --resume OOM` when `o` is your shell alias): ```bash theme={null} opensre --resume ``` In a TTY, bare `/resume` opens an interactive picker of recent sessions. New turns after a resume append to that same session file — `/resume` does not create a duplicate entry in `/sessions`. To resume a specific branch point, append an entry ID prefix after the session ID: ``` /resume 9b2e4f7a:abc123 ``` When the branch contains a compaction entry, replay inserts the saved summary before the kept messages so the assistant has the compacted context. OpenSRE restores the conversation, shows the past exchanges, then lets you continue: ``` resumed session 9b2e4f7a · why is CPU spiking on prod-api (14 messages) ─── conversation history ──────────────────────────────────── ❯ why is CPU spiking on prod-api? ● assistant The root cause is a connection pool leak in the orders service... ───────────────────────────────────────────────────────────── ``` **What gets restored:** | | Restored by `/resume` | | - | - | | Full conversation context (what you asked, what the assistant said) | Yes | | Infra context (service name, cluster, region learned mid-session) | Yes | | Trust mode, reasoning effort | No — these are per-session preferences | You can also search by name instead of ID: ``` /resume redis # resumes the session whose name contains "redis" /resume cpu spike # matches by name substring ``` If the prefix matches more than one session, OpenSRE asks you to be more specific. *** ## When you exit When you leave the REPL, OpenSRE prints your session ID so you can resume later: | Exit path | Behavior | | - | - | | `/exit` or `/quit` | Prints `Resume this session with:`, `/resume `, and `opensre --resume `, then goodbye | | **Ctrl+C** twice within 2 seconds | Same resume hint, then exits | | **Ctrl+D** (EOF) | Same resume hint when no dispatch is running | Example: ``` Resume this session with: /resume 3f8a1c2d-… opensre --resume 3f8a1c2d-… Goodbye! ``` Copy the `/resume …` or `opensre --resume …` line from the terminal, or find the session later with `/sessions`. *** ## Starting fresh with context — `/new` `/new` closes the current session and opens a new one, but carries the conversation forward: ``` /new ``` ``` new session started — conversation context carried forward. 14 messages in context · type to continue ``` Use `/new` after a long session to keep `/sessions` tidy without losing your place in the conversation. *** ## Compacting long sessions — `/compact` OpenSRE replays earlier turns of the session to the model in order, including the tool calls it made and what they returned (long tool output is shortened to its start and end). A follow-up such as "rerun the one that failed" can use a run ID from an earlier answer without fetching it again. When that history grows past its token budget, OpenSRE compacts it before the next turn: the model writes a summary of the older turns, and the most recent turns stay word for word with their tool results. `/compact` does the same on demand and keeps only your latest exchange: ``` /compact ``` The compaction is saved in the session file, so `/resume` restores exactly the summary and the turns that were kept. Run `/context` to see how much of the budget the history uses and what else the next model call carries. | Variable | Default | Effect | | - | - | - | | `OPENSRE_HISTORY_TOKEN_BUDGET` | `48000` | Estimated tokens of history before automatic compaction | | `OPENSRE_HISTORY_TOOL_RESULT_CHARS` | `8000` | Characters of each tool result kept in history | | `OPENSRE_LLM_COMPACTION` | `1` | Set to `0` to summarize without a model call | | `OPENSRE_STRUCTURED_HISTORY` | `1` | Set to `0` to send earlier turns as a single text block (the previous behavior) | **`/new` vs `/clear`:** | Command | What it does | | - | - | | `/clear` | Clears the terminal screen only — session and context unchanged | | `/new` | Opens a new session file; conversation context carried forward | *** ## What OpenSRE saves Each session records: * When it started and ended, and how long it ran * Every message you sent and every response from the assistant * Tool calls, tool updates, tool results, model/tool changes, labels, custom messages, and compaction summaries * Infra context discovered while working (service names, cluster names, regions) * Turn counts Sessions are stored as files under `~/.opensre/sessions/`. Each file is plain text and human-readable. ## Inspecting a turn's decisions Local session traces record the model's original closing, the reason a stop was accepted or rejected, the plan and pending questions, and the displayed and saved answers separately. These diagnostic records stay out of resumed conversation context. Known credential patterns are redacted; ordinary conversation text is retained. Run `/sessions` to find the session ID, then use `jq` in your terminal. Replace `3f8a1c2d` with the ID prefix shown for your session: ```bash theme={null} jq 'select(.type == "trace_span" and .span_kind == "decision")' \ ~/.opensre/sessions/3f8a1c2d*.jsonl ``` Look for `goal_review` and `conclusion` to understand stopping, `agent_finished` for the original closing, `response_composition` for display filtering, and `response_displayed` / `assistant_persisted` for the two outputs. Rejected closing attempts appear as `model_conclusion` records. The records include the prompt turn ID when available; `assistant_persisted.message_id` identifies the saved message. These records are added on new runs; older traces cannot recover text that was never recorded. *** ## Privacy and what is not saved * **Secrets or credentials** — known token shapes are redacted from the prompt log by default (set `OPENSRE_PROMPT_LOG_REDACT=0` to store raw text instead) * **Token usage** — tracked in `~/.opensre/prompt_log.jsonl` when prompt logging is enabled See [Interactive Shell Privacy](https://opensre.com/docs/platform/interactive-shell-privacy.md) for redaction controls. *** ## Related * [Interactive Shell Privacy](https://opensre.com/docs/platform/interactive-shell-privacy.md) — command history, prompt/response logging, and redaction settings * [Environment variables](https://opensre.com/docs/install/environment-variables.md) — full prompt/response logging configuration This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.