Investigations & the agent

Work in Codepanion is organised into investigations — a chat thread where you ask a question and the agent works toward an answer. It responds by streaming tokens live, and uses tools as it goes:

• search_codeSemantic (vector) search over your indexed source — finds the relevant files and lines by meaning, not just keywords.
• search_textExact keyword search over the same indexed source — file paths included. The right tool when the agent has an identifier, error message, or config key in hand and needs every place it appears verbatim.
• query_databaseRuns a read-only SELECT against your connected database and returns the rows.
• read_codepanionPulls in the relevant .codepanion context file when it needs domain knowledge.
• search_issuesLooks up tickets in a connected issue tracker (see integrations).

You watch each step happen as it runs — nothing is hidden. The agent's findings cite the exact files and lines it read, so you can verify the answer rather than take it on faith.

Any message in an investigation can be forked into a new path. Hit the branch action on an earlier message and the conversation carries on from that point, leaving the original intact.

It's the "what if I'd asked it differently" button: explore an alternative line of reasoning without losing the thread you already have. When a message has more than one continuation, a Branch N of M switcher lets you move between them.

Every tool call the agent makes is rendered as a short, plain-language step — "searched your code for refund handling," "ran a query against orders" — so you can follow the reasoning without reading raw JSON.

Need the underlying detail? Each step has a Show details disclosure that reveals the exact tool input and output — the query that ran, the files that matched, the rows returned. Friendly by default, fully auditable on demand.

Each investigation shows a live cost meter in the chat header — the tokens used so far and the running cost in dollars, updated after every turn. No surprises at the end of the month: you can see exactly what a given investigation is costing as it happens.

Usage is metered per tenant, so admins can also review spend across the whole account in the hub.

Any investigation can be exported as a Markdown file — the full conversation, the tool steps, and the code citations included. It's the quickest way to drop a finding into a ticket, a postmortem, or a Slack thread, or to hand a teammate the whole trail of how an answer was reached.

Some context is true across every investigation — a flaky downstream service, a quirk in how a legacy table is populated, a workaround support already knows about. You can pin these as known issues for your tenant, and the agent factors them in automatically on every investigation.

It saves re-explaining the same caveat each time, and keeps the whole team's answers consistent with what you already know.