Help & Frequently Asked Questions

Releap is codebase intelligence and workflow software for product, support, customer success, engineering leadership, and operators. Teams ask natural-language questions about how their product works, get cited answers from code and connected product context, then turn those answers into tickets, epics, Knowledge Base pages, release communications, and opt-in draft PRs.

Anyone on a software team who needs to understand what engineering built without reading code line by line. The primary users are product managers, customer success engineers, operations leads, engineering managers, and founders. Baseline Releap is read-only; optional branch-writing workflows require explicit workspace admin enablement and a separate GitHub permission grant.

Releap is not a general-purpose autocomplete tool or an IDE-native pair programmer. It starts from product understanding and organizational context, then helps teams plan, document, communicate, and, when explicitly enabled, generate reviewable draft PRs from Releap tickets. It does not access your issue tracker by default.

No. Releap is specifically designed for non-engineers. You ask questions the way you would ask a senior engineer: "How does the password reset flow work?", "Where is rate limiting enforced?", "What happens when a payment fails?" Releap answers in plain language with citations you can verify.

Releap currently supports GitHub via the Releap GitHub App. GitLab, Bitbucket, and Azure DevOps are not currently supported but are on the roadmap.

Releap uses AST-aware chunking via tree-sitter, which supports all major programming languages including Go, Python, TypeScript, JavaScript, Ruby, Rust, Java, C, C++, and more. It also indexes configuration files, manifests, and documentation files (Markdown, YAML, JSON, TOML). Binary files and files over the size limit are skipped. Files in your `.gitignore` are not indexed.

Yes. You can connect a monorepo and use Releap's application scoping to organize it into named product surfaces. Queries can be scoped to a specific application, branch, or the full repo.

Indexing time depends on repo size. A typical repo of 50,000 lines of code indexes in 2–5 minutes. Large monorepos with hundreds of thousands of lines may take 15–30 minutes for the initial index. Subsequent incremental reindexes on push events are significantly faster — usually under a minute for typical changesets.

Yes. Releap reindexes automatically on GitHub push events via webhook. When your team pushes code, Releap updates the affected files incrementally. You can also trigger a manual full reindex from the workspace settings.

In addition to GitHub repositories, Releap can index Confluence spaces, Jira projects, and Aha! features and epics. These are indexed alongside your code and retrievable in the same query interface with the same citation format.

Yes. Planning Mode is a guided epic planning flow built into the Releap chat interface. Start by asking questions about your codebase, then switch to Planning Mode to generate a structured epic with child stories, acceptance criteria, and code references grounded in the actual code. Releap asks clarifying questions based on what it finds in the codebase, validates your spec against the current implementation, and surfaces the questions engineers are likely to ask. Export the finished epic directly to GitHub Issues, Linear, Jira, or Aha!

Growth gives you SSO, BYO LLM, BYO embedder, ticket-to-draft-PR workflows, and a full audit log without a custom contract — just self-serve annual billing. Enterprise is for organizations that need a signed Master Services Agreement, a negotiated SLA with financial penalties, a custom Data Processing Addendum, multi-workspace deployment, custom retention, or dedicated security-review support. If you're not sure which is right, contact us and we'll help.

Every new workspace gets a 30-day Pro trial with no credit card required. During the trial you get full Pro features: web UI, MCP access, AI-generated thread titles, ticket generation, Knowledge Base pages, and included monthly AI usage. At day 30, if you haven't subscribed, your workspace moves to read-only — your indexed repos and chat history are retained and fully restored when you subscribe.

Nothing is deleted. Your workspace moves to read-only: no new queries, no new indexing, no push-triggered reindexes. Your data waits for you. Subscribe at any point to restore full access immediately.

The current pricing model is based on monthly AI usage, not separate query credits. A message that triggers retrieval and an LLM response consumes the shared workspace pool, as do other AI actions such as ticket drafting, planning, summaries, KB generation, release-note drafting, verification, and draft-PR workflows.

Each plan includes a shared monthly AI usage pool for the workspace. AI actions such as chat answers, ticket drafting, planning, summaries, KB generation, release-note drafting, verification, and ticket-to-draft-PR workflows draw from the same pool; reading existing answers, tickets, KB pages, release notes, and settings does not. When a workspace needs more, it can purchase 1M-token usage blocks: $10 on Pro, $8 on Business, $6 on Growth, and contracted terms on Enterprise.

Pro is for individual contributors — per-seat billing at $25/seat/month, web UI and MCP access, 15 repos, GitHub Issues export, Knowledge Base pages, and Standard monthly AI usage per seat. Business is for teams — flat-rate $299/month, unlimited seats, Slack bot for shared-channel querying, a larger shared AI usage pool, 35 repos, all tracker exports, Confluence and Jira ingest, release communications, and admin prompt controls.

Yes. Seats are added immediately and prorated to your next billing date. Removing a seat applies at the next billing cycle.

Yes. Pro annual is $240/seat/year ($20/seat/month equivalent — 20% off monthly). Business annual is $2,868/year ($239/month equivalent — 20% off monthly). Growth annual is $6,708/year ($559/month equivalent — 20% off monthly). Annual plans are charged upfront.

Yes. Upgrades apply immediately. Downgrades apply at the end of your current billing period.

Enterprise is a custom annual contract for organizations that require a signed MSA, negotiated SLA with financial penalties, custom Data Processing Addendum, multi-workspace deployment, or dedicated security-review support. Enterprise includes everything in Growth — SSO, BYO LLM, BYO embedder, ticket-to-draft-PR workflows, full audit log, Security Audit Mode — plus unlimited repos, contracted AI usage, configurable retention (1-7 years), SOC 2 report under NDA, and a dedicated account manager. If you only need SSO, BYO, and draft-PR workflows without a custom contract, see Growth instead. Contact us to discuss Enterprise.

You can cancel from workspace settings → billing at any time. Your subscription remains active through the end of the current billing period. No data is deleted on cancellation — your workspace moves to read-only and your data is retained.

Releap follows the delivery loop from understanding to planning to implementation to verification to documentation to release communication. Teams start in chat with cited answers, turn findings into tickets or epics, export to trackers, hand work to developers or agents, verify acceptance criteria, capture durable knowledge in the Knowledge Base, and draft audience-specific release notes from the same grounded context.

Chat supports multi-turn investigations, application and branch scoping, code and product-source retrieval, cited answers, thread history, summaries, ticket signals, and action hints. A chat thread can become a Releap ticket, an epic planning session, a KB page, a release-note input, or an agent handoff without losing the evidence that led to the answer.

Planning Mode turns exploratory Q&A into structured delivery work. Releap can draft an epic, child stories, acceptance criteria, implementation notes, code references, and review questions grounded in the current codebase. The finished plan can be exported to GitHub Issues, Linear, Jira, or Aha! depending on your plan.

A Releap ticket can include title, description, ticket type, status, priority, story points, due date, acceptance criteria, code references, comments, linked external tracker IDs, and a suggested fix prompt. Tickets can be searched, compared, exported, updated, and used as the source for draft PR work when branch-writing is enabled.

Release communications use tickets, milestones, code context, commit summaries, KB content, and approvals to draft audience-specific notes. Releap tracks review state, publication destinations, delivery records, and published communication chunks so future answers can include what was actually communicated to customers or internal teams.

KB pages are generated from cited answers and tickets, edited by the team, retrieved in chat, searched directly, and flagged when cited source files change. This means support, product, and engineering can keep durable product knowledge close to the code that makes it true.

Releap can record acceptance-criteria verification, linked PR metadata, commit SHAs, and local or CI test results. MCP clients and the VS Code extension can report verification outcomes back to Releap, while Releap itself does not run arbitrary repository commands from MCP.

The MCP server lets Claude Code, Cursor, Cline, Codex, and other MCP-capable tools use Releap as a grounded context and workflow layer. Agents can query the codebase, retrieve cited chunks, list visible repos, create and compare ticket drafts, generate epics, update stories, search and update tickets, record verification, update KB articles, update release notes, and coordinate opt-in branch work.

MCP gives the agent product memory it normally lacks: current code citations, tickets, acceptance criteria, KB pages, release-note history, and workspace rules. Instead of pasting context manually, the agent can fetch the relevant Releap context on demand, ask follow-up questions, propose ticket or documentation updates, and record what it verified.

Only through the explicit Releap Code workflow. The workspace admin must enable branch-writing and grant the separate GitHub content-write scope. Releap applies tickets only to auto-generated or user-designated non-default branches and opens draft PRs for review. It never targets the default branch, merges PRs, or changes repository settings.

Core tools include query_codebase, list_repos, read_chunk, releap_create_ticket, releap_compare_tickets, generate_epic, update_story, run_review, export_epic, search_tickets, update_ticket, apply_ticket_to_branch, create_pr, verify_acceptance_criteria, run_tests, update_kb_article, and update_release_note. Availability depends on workspace plan, permissions, and enabled integrations.

MCP access uses workspace-scoped API keys or OAuth-connected agent tokens. Calls are authenticated, audited, plan-limited, and governed by the same repository visibility, application access, sensitivity handling, BYO model configuration, and monthly AI usage policies as the web app.

The VS Code extension brings Releap into the editor with a personal working queue, ticket details, cited file links, native @releap chat commands, agent handoff prompts, local acceptance checks, Releap Code controls, workspace and AI usage status, and MCP setup snippets for Claude Code, Codex, Cursor, or another MCP client.

No. The extension helps you copy MCP configuration and keeps the queue, chat, verification, and handoff workflow close to the code. MCP-capable agents connect directly to the Releap MCP endpoint using their own configured client.

Yes. Releap connects via the Releap GitHub App using read-only installation tokens. Private repos are fully supported and never publicly exposed.

No. Releap uses a default-deny model. When you install the GitHub App, no repositories are visible to Releap until a workspace admin explicitly grants visibility to each one. You control exactly what Releap can and cannot see.

Releap exposes a Model Context Protocol (MCP) endpoint that allows external coding agents such as Claude Code, Cursor, Cline, Codex, and others to query your codebase and product memory using workspace-scoped credentials. This lets your coding agent pull cited context, ticket details, KB content, release-note history, and action hints on demand without leaving its interface.

Releap generates ticket drafts on demand — it does not create tickets automatically. When a query uncovers work to do, you click to generate a ticket draft with title, description, acceptance criteria, code references, and a fix prompt. You review and export it to your tracker (GitHub Issues, Linear, Jira, or Aha!) with one click.

Yes, when explicitly enabled by a workspace administrator. A Releap ticket with code references can be applied to an auto-generated or user-designated non-default branch, then opened as a draft pull request for engineering review. This requires a separate opt-in GitHub content-write permission. Releap never targets the repository default branch, never merges pull requests, and never modifies repository settings.

The Knowledge Base is a Releap-owned internal documentation surface. Teams can create spaces, pages, and sections from cited answers or tickets, edit them collaboratively, retrieve them in chat, search them directly, and see staleness signals when cited source files change.

Release communications let teams draft audience-specific notes from tickets, milestones, code context, and KB content. Releap tracks approvals, destinations, publication records, and published communication chunks so future answers can include what you told customers.

GitHub Issues (Pro and above), Linear, Jira, and Aha! (Business and above). Ticket exports are idempotent — exporting the same draft twice updates the existing issue rather than creating a duplicate.

Yes, on Growth and Enterprise plans. Workspace administrators can provision Security Audit Mode grants for authorized security assessors. In this mode, designated users can query specific repositories for security-sensitive content that is normally suppressed by Releap's sensitivity classifier — authentication logic, cryptographic handling, and security documentation. All audit-mode queries are logged with mandatory full retention. Grants are time-bounded (maximum 30 days), repo-scoped, and expire automatically. Self-escalation is not possible.