Live betaCtrlB Recall

Keep your agents working beyond the context window

CtrlB Recall keeps commands, errors, decisions and tool results searchable across sessions. Your agents find the context they need to continue long tasks, even after compaction.

Illustrative example

Agent context

25%
SummaryFixed the retry tests in refunds. Next: update the docs.
  1. #38PromptKeep refund retries under 3.
  2. #41bashpytest tests/refunds -k retry
  3. #42ResultFAILED test_retry.py::test_cap
  4. #57Editsrc/refunds.py
  5. #90PromptWhich test failed before the fix?
  6. #91Searchsearch_sessions("retry.py")
  7. #92AnswerFound at #42: test_cap in test_retry.py failed.Exact row

CtrlB Recall

4 rows kept
  1. #38PromptKeep refund retries under 3.
  2. #41bashpytest tests/refunds -k retry
  3. #42ResultFAILED test_retry.py::test_capHit
  4. #57Editsrc/refunds.py
Compaction removes rows from the context. CtrlB Recall keeps every row, and the agent finds it with one search.
  • 81 to 94%Detail questions right with CtrlB Recall
  • 22 to 35%Right with the compaction summary only
  • 0 model callsTo build memory. Searchable in 5.5 s
  • 1.4 s vs 189 sSearch at 100× the record, against grep
After compaction

Recover what happened, not just the summary

A compaction summary keeps the outcome. CtrlB Recall keeps the steps that led to it.

Illustrative example

After compaction

“The previous test failed.”

The summary keeps the outcome. The command, the error and the fix are gone.

Kept: 1 line

With CtrlB Recall

Command
pytest tests/auth/test_token.py -x
Error
FAILED test_refresh: expected 401, got 200
Change
auth/token.py: check expiry first

Source: session 7f3a2c, rows 4,182 to 4,190

Memory that decides when it writes has to guess which details will matter. CtrlB Recall keeps every row, so the agent decides at read time, when it knows the question.

How CtrlB Recall works

Three steps, and your agent runs all of them through hooks and tools.

  1. 01

    Capture

    Save messages, tool calls and results from connected agent sessions.

  2. 02

    Keep

    Keep the original record while the agent's context is compacted.

  3. 03

    Recall

    Find earlier commands, errors and decisions through the agent's tools.

Illustrative example
  1. Your agent

    Claude Code or Codex runs as usual.

    Subagents too.

  2. Hooks upload

    At the end of each turn, before compaction and at session end.

    Secrets are redacted first.

  3. Your bucket

    One row per message, tool call and tool result.

    Open Parquet in S3, GCS or Azure Blob.

  4. Recall tools

    The agent searches the record over MCP.

    search_sessionsfind_commandnext_actionsfetch_records
Search before you answer

At session start, the agent is told: search before you answer.

At session start, the plugin adds one line to the agent’s context: search the record before you answer about earlier work. Without that line, GPT-6 Luna made no recall call in 90% of runs. Hooks redact text that looks like a secret before upload, but the patterns are not a guarantee.

What your agents can ask

Four MCP tools read the record, and each reply stops at 12,000 characters.

  • search_sessions

    Find a word across sessions

    It finds a whole word in messages, commands and results, in current and past sessions. It returns 10 hits by default, 25 at most.

    search_sessions(term="refunds.py")

    You get: Hits with session, row number and a 300-character snippet.

  • find_command

    Find a command you ran before

    It returns every past tool call whose command contains your text, newest first. It returns 10 by default, 25 at most.

    find_command(command="pytest tests/refunds")

    You get: The total, then session, row and command for each match.

  • next_actions

    See what happened next

    It returns the result of one tool call, then the next three calls and their results. Text and thinking rows are skipped.

    next_actions(session_id="7f3a91c2", seq=41)

    You get: What the agent did after a failure, and whether it worked.

  • fetch_records

    Open the rows around a hit

    It returns the rows of a range in order. A range holds 50 rows at most, and each body is cut at 4,000 characters.

    fetch_records(session_id="7f3a91c2", seq_from=38, seq_to=42)

    You get: The exact rows, with the full length of any cut body.

Recover more details after compaction

Four models, two families, asked at real compaction points.

Four engineers’ sessions, 100 detail questions per model, with a one-line instruction to search.

Detail questions right: summary only against CtrlB Recall

Detail recall, correct answers (%)
ModelSummary onlyCtrlB Recall
Claude Haiku 5.523%85%
Claude Sonnet 5.522%89%
GPT-6 Luna29%81%
GPT-6.1 Sol35%94%
Internal evaluation. Four models in two families, 100 detail questions per model, one run per question. The summary stays in context in both arms. The Recall arm also has the full session record and a one-line instruction to search earlier work.

Better recall doesn't always need a bigger model

Claude Haiku with CtrlB Recall answered more than Claude Sonnet with only the summary.

Same 100 detail questions. Haiku 5.5 with Recall costs about 1/15 as much per question as Sonnet 5.5.
  • 70 to 72%Cross-session lookup with CtrlB Recall, small models. Summary only answered none.60 questions per family, no session named.
  • -2 to +6 ptsRecall change when the summary is dropped, with CtrlB Recall and the search instruction on. Each answer cost less.Compaction pauses a Claude Code session for a median of 61 s.

With CtrlB Recall on both sides, the larger model is level or ahead. Haiku 5.5 trails Sonnet 5.5 by 4 points, not significant. GPT-6 Luna trails GPT-6.1 Sol by 13.

The limit: four engineers at one company over six weeks, and a model judged the answers.

Search matches exact words. We did not test search by meaning.

Compared with extracted memory

How CtrlB Recall compares with Mem0

Mem0 keeps short extracted memories. CtrlB Recall keeps the full record and searches it.

Illustrative example

Turn 90 asks

Which test failed before the fix?

Compaction summary

Decides at compaction

Fixed the retry tests in refunds. Next: update the docs.

Not in the summary

29%detail recall

Extracted memory

Decides when it writes

  • Prefers pytest
  • Refunds live in src/refunds.py
  • Retries capped at 3
Never extracted

37%detail recall

CtrlB Recall

Decides when your agent asks

  1. #41bashpytest tests/refunds -k retry
  2. #42ResultFAILED test_retry.py::test_cap
  3. #57Editsrc/refunds.py
#42: test_cap failed

81%detail recall

One model (GPT-6 Luna) and the same 100 detail questions for all three, asked at real compaction points.
AccuracyDetail questions answered
CostModel calls to build memory
SpeedTime until searchable

Internal evaluation against open-source Mem0 2.2.1. Accuracy: GPT-6 Luna and GPT-6.1 Sol, 100 detail questions each (Mem0 was not run with Claude). Build: one engineer's 538 Claude Code sessions. Cost per question answered was the same, $0.003.

Find answers across your team's agent sessions

Recover past investigations, commands and fixes without knowing which session holds them.

  • Which command fixed this error?
  • Why did we change this configuration?
  • What did the other agent already try?

Search the sessions you're authorized to access.

  • You see your own sessions by default. The sessions.read_all permission shows everyone’s.
  • Rows are open Parquet files in your S3, GCS or Azure Blob bucket.
  • Search stays fast as the record grows.
Median search time of CtrlB against grep over the files, 120 calls at each size.
Record sizeCtrlBgrep over files
1× (170k rows)147 ms450 ms
10× (1.7M rows)494 ms4.1 s
100× (17M rows)1.05 s165 to 408 s (9 of 120 calls finished)

At 100×, the files are larger than the test machine’s memory, and grep finished 9 of 120 calls. On those nine calls, CtrlB took 1.4 s and grep took 189 s.

Sample data
retry.pyShow: Top-level
Sample sessions table with four agent sessions
SessionUserHarnessLast activityTurnsTool callsErrors
Keep refund retries under 3.7f3a91c2 · billing-api · fix/retry-cappriya@acme.devClaude Code10:42461122
Add idempotency keys to the refund endpointc41e08b7 · billing-api · maindan@acme.devCodex10:1531740
Why does the invoice export time out?9b20d4ae · invoices · perf/exportpriya@acme.devClaude Code09:5818401
Write the migration for refund_attemptse83f5c10 · billing-api · fix/retry-capsam@acme.devClaude Code09:3122510
4 sessions

Works with your agents and Relay

The CtrlB plugin connects Claude Code and Codex.

  • Claude Code logoClaude Code
  • Codex logoCodex

Your agent history, in your own cloud

CtrlB Recall stores session records in your own bucket as open Parquet. Connect your agents through integrations, MCP or the API.

AI Observability pricing

Start with the agents you run today.

Tracing, evaluations, investigations, and Guardrails in one plan.

  • Production traces and model cost tracking
  • Scheduled investigations and prompt comparisons
  • Tool policies with Guardrails
  • OpenTelemetry, your own bucket and your own model key

AI Observability

From $499 / month

Platform fee. Usage is quoted separately for your workload.

Discuss your workload

We agree usage costs before you start. Your providers bill model and bucket costs directly.

Need a larger deployment?Enterprise pricing
FAQ

Common questions

What does CtrlB Recall keep?

It keeps one row for each message, tool call and tool result. Each row has the role, the shell command, the time and the session. Subagents are included. Nothing is updated or deleted.

How is it different from extracted memory?

Extracted memory decides what to keep when it writes. CtrlB Recall keeps every row, and your agent decides what matters when it asks. With GPT-6 Luna on 100 detail questions, extracted memory scored 37% and Recall scored 81%. Against open-source Mem0 2.2.1, Recall answered 81 to 94% of detail questions and Mem0 answered 37 to 44%.

How is it different from compaction?

Compaction writes a summary and drops the rows. In our Claude Code sessions, 6% of exact answers were still in the context after compaction. Exact means a path, number, error line or command. Recall keeps the rows, so the agent can find them.

Where does my data live?

Rows are open Parquet files in your S3, GCS or Azure Blob bucket, and SQL can query them. Hooks redact text that looks like a secret before upload. The redaction patterns match common key and token shapes and are not a guarantee.

Who can see whose sessions?

You see your own sessions by default. The sessions.read_all permission shows every user's sessions. The sessions.write permission lets your agents upload them.

What does it cost, and how fast is it?

At 100× the record, CtrlB search took a median of 1.05 s. Grep over the files finished 9 of 120 calls. Haiku 5.5 with Recall costs about 1/15 as much per question as Sonnet 5.5. Pricing is on this page.

Do I need Relay?

No. CtrlB Recall works on its own. Relay builds its runs from Recall records, so agents in a Relay run need sessions.write.

Which agents does it work with?

Claude Code and Codex, through the CtrlB plugin and hooks. The hooks upload at the end of each turn, before compaction and at session end.

Live beta

See CtrlB Recall with your agents

Live beta. From $499/month, plus usage.