Research report · data-report
Whole-codebase work: a census of 143 listings
14 of 143 listings answer yes for “Whole-codebase work”. Counted from our own datasheet, with the query that reproduces every figure.
On this page
Whole-codebase work, across 143 settled listings
Of the 143 Published listings whose datasheet settles “Whole-codebase work”, the largest group is “Does not apply” at 76 (53%). Measured 2026-10-03.
Explore the matching agents
Whole-codebase work · Datasheet criterion codebase_scale · Report counted .
Current listings as of ; this set may differ from the report's original sample. Includes “Yes” and “Qualified” answers only. Read the conditions on qualified answers before treating them as a match. “No”, “Not applicable” and “Not established” answers are excluded.
View 53 current matching records
- AgentsKit.js Qualified
Multi-file edits and refactors are claimed for the coder persona over a filesystem scoped to one basePath; no repository indexing or repo-wide migration capability is documented.
Source retrieved 2026-09-09
- Agno Qualified
Repository-wide migrations are a named pattern, but the docs expect unresolved cases to be reported rather than fully landed.
- Aider Qualified
A ranked map of the whole git repo gives it repository-wide context, but the vendor warns performance is not optimised for very large monorepos and only one repo can be open at a time.
Source retrieved 2026-09-30
- Aisera Agent Composer (Agent Studio) Qualified
The granularity the vendor claims is lines and files, not repositories: conversion works on lines of code and codebase support is explanation and summarisation, with no repo-wide migration claim.
Source retrieved 2026-09-03
- Antigravity CLI Yes
It indexes the trusted workspace and the docs direct large refactors and repository-wide sweeps to concurrent background subagents rather than one serial thread.
Source retrieved 2026-09-10
- Augment Code Yes
Published workflows cover module-by-module legacy migration and epic decomposition across repositories.
Source retrieved 2026-09-02
- AutoGen Qualified
FileSurfer navigates directories and reads local files while the executors run commands in a work_dir, so file-by-file work is possible; no repository indexing or repo-wide migration is documented.
- Bitsclan IT Solutions Qualified
They take on modernisation of existing systems, but the published scope is infrastructure analysis and a human-authored plan, with no claim of carrying a change across an entire repository.
Source retrieved 2026-09-11
- Bolt Qualified
It works across a whole imported repository and claims far larger context, yet the docs still advise splitting files over 700 lines and clearing context between tasks.
Source retrieved 2026-09-11
- Claude Agent SDK Qualified
Repo-scale work is handled by fanning out subagents and by automatic compaction; no whole-repo index is claimed.
Source retrieved 2026-09-02
- Claude Code Yes
Dynamic workflows run many subagents over a repository for audits and large migrations; the vendor's own example is a 500-file migration.
Source retrieved 2026-09-02
- Cline Yes
Repository-wide migrations are a stated use case, though the vendor is explicit that it never indexes a repo and instead uses partial-file analysis and summarisation.
Source retrieved 2026-09-12
- CodeRabbit Qualified
Reviews reach beyond one repo through linked-repository analysis, but the number of linked repos is capped per plan (1 on Essentials) and automatic linking needs Team or above.
Source retrieved 2026-09-15
- Codex CLI Qualified
Documented for repository-wide exploration and multi-step plans; the vendor states no repo-wide migration guarantee.
Source retrieved 2026-09-02
- Cursor Yes
Cloud agents work across multiple repositories in one task and raise a pull request in each repo they touch.
Source retrieved 2026-09-02
- Devin Yes
Migrations spanning many files, modules or whole repositories are the stated use for fanning work out to parallel sessions.
Source retrieved 2026-09-02
- Devin Desktop Yes
Retrieval is delegated to the SWE-grep Fast Context subagent; Codemaps chart execution flow across a repository and Teams/Enterprise can index remote repos server-side.
Source retrieved 2026-09-03
- Dust Qualified
It can ingest a repository's default-branch code for search, refreshed every eight hours, but no documented ability to carry a change across that repository.
Source retrieved 2026-09-09
- Factory Yes
Missions decompose multi-feature work into milestones and are documented for brownfield migrations.
Source retrieved 2026-09-02
- Gemini Code Assist Yes
Local codebase awareness is quoted at a 1,000,000 token context window; Enterprise adds code customization indexed over private GitHub, GitLab and Bitbucket repositories.
Source retrieved 2026-09-02
- Genspark Qualified
Genspark Code lists 'Existing code' as one of five project categories, and GenTeam local agents can be pointed at a repository directory, but repo-wide migration scale is never described.
Source retrieved 2026-09-15
- GitHub Copilot Qualified
Works across a whole repository, including Java and .NET modernisation, but one run is confined to one repository and one branch.
Source retrieved 2026-09-02
- Glean Qualified
Code Writer is scoped to a small targeted change in one configured repository and branch, not a repository-wide migration.
Source retrieved 2026-09-02
- Google Agent Development Kit (ADK) Qualified
The documented route across an existing repository is Agents CLI skills driving your own coding agent through an incremental migration; ADK itself carries no repository-wide refactor engine.
Source retrieved 2026-09-15
- goose Qualified
The Analyze extension traces symbols and call graphs across directories, but results over 50,000 characters are withheld unless forced, and the docs advise narrowing scope or delegating to a subagent, so repository-wide work is chunked rather than indexed.
Source retrieved 2026-09-16
- Junie Yes
The vendor claims repository-wide edits rather than single-file assistance, and markets long-running migrations specifically as the case its asynchronous remote mode exists for.
Source retrieved 2026-09-18
- Kilo Code Yes
Indexes the repository and runs multi-step repo-wide work such as migrations.
Source retrieved 2026-09-02
- LangChain Qualified
Engine reads a whole connected repository to ground fixes and dcode carries context across sessions, but no page claims a repository-wide migration or multi-file refactor as a supported job.
Source retrieved 2026-09-18
- Letta Qualified
Subagents read across large repositories and memory persists project knowledge, but no repo-wide migration claim.
Source retrieved 2026-09-02
- Manus Qualified
Works across the whole project it builds and keeps a two-way sync with a private GitHub repo, pulling main before each change; carrying a migration across a pre-existing repository is not described anywhere I read.
Source retrieved 2026-09-19
- Mastra Qualified
Agents you build get file tools over a mounted tree (grep and structural edits), but Mastra claims no repository-wide migration capability of its own.
- Microsoft Agent Framework Qualified
The Harness targets long multi-step work with context compaction, file memory and todo tracking, and shell tools default to working-directory confinement; the vendor documents no repository indexing or repo-wide migration.
- Muse Code Qualified
It splits a large job across parallel subagents with optional Git worktree isolation, which fits a repository-wide migration, but the vendor documents no repository indexing and caps a tree at eight concurrent agents by default.
Source retrieved 2026-09-20
- Octacer Qualified
Large migrations are sold as sequenced, risk-sliced modernisation projects run by their team, not as a repository-wide change an agent performs in one pass.
Source retrieved 2026-09-26
- OpenAI Agents SDK (Python) Qualified
A sandbox manifest can stage a whole local directory or git repository into the workspace, so repository-wide work is possible, but the vendor documents staging and sessions rather than any migration or indexing capability.
Source retrieved 2026-09-20
- OpenClaw Qualified
Managed worktrees give a task its own branch and checkout with snapshots and cleanup, which supports repository-wide work, but the vendor never claims a migration carried across a whole codebase.
Source retrieved 2026-09-21
- OpenCode Qualified
It loads language servers and has a read-only Explore subagent for whole-repo search, but the vendor never claims repository-wide migrations.
Source retrieved 2026-09-21
- OpenHands Qualified
Repository-wide migrations are documented as prompts the agent handles, but the Large Codebase SDK that targets big repos is an Enterprise-only line on the pricing table.
Source retrieved 2026-09-21
- Perplexity Qualified
The agent clones a whole repository and works in it through git and gh, but Perplexity never claims a migration carried across a repo, and the connector is labelled preview.
Source retrieved 2026-09-02
- PolyAI Qualified
Wren applies one change across every part of an agent project at once, but it is scoped to the project you are in and cannot reach other projects.
- Pydantic AI Qualified
The Harness is explicitly aimed at long-running repository work with repo context and context management, but the framework supplies the machinery rather than a measured guarantee about repository-wide migrations.
Source retrieved 2026-09-21
- Replit Agent Qualified
Max Mode is positioned for larger codebases and tasks can carry dependencies, but no repository-wide migration guarantee or size limit is published.
Source retrieved 2026-09-22
- Salesforce Agentforce Qualified
Parallel sub-agents coordinate multi-file, cross-layer changes in isolated Git worktrees, but the docs warn that context truncates, advise preferring specific files over entire folders, and claim no repository-wide index.
Source retrieved 2026-09-03
- Sim Yes
Create PR clones the whole repository into a sandbox with full read, shell, edit and git access.
- SkaleAgents Qualified
An application-wide review maps related source files that the editor sends, but there is no repository indexing, and scheduled audits reach only one file in a public GitHub repository.
Source retrieved 2026-10-03
- Tale Qualified
Each project agent keeps a persistent sandbox workspace and may clone repositories, but the vendor documents no repository-wide indexing or migration capability and caps runs by time and sandbox capacity.
Source retrieved 2026-10-03
- TecAdRise Qualified
Migrations are sold as scoped human engineering projects, such as porting a legacy MetaTrader strategy to a newer platform, not as an agent that carries a change across a client repository.
Source retrieved 2026-09-12
- Vecbase Qualified
A repository link or nearby files can be supplied as context; the vendor claims no whole-repo indexing and no repo-wide migration.
- Warp Yes
Codebase Context indexes the Git-tracked repo (and other indexed repos) so agents can carry repo-wide changes; index size caps vary by plan, at least 5,000 files per codebase on every plan.
Source retrieved 2026-09-03
- Zed Qualified
The agent searches the repository with grep and find_path and can spawn subagents for parallel investigation, but Zed asks you to supply context explicitly and documents no repository-wide index.
Source retrieved 2026-09-29
- Zencoder Qualified
Agents work across a whole repository on any plan, but the cross-repository indexing that carries a change through several services starts at Pro Plus and is admin-configured.
Source retrieved 2026-09-09
- zot Qualified
It works across a whole working directory with recursive glob search and a canned whole-project reading pass, but there is no repository index; large work depends on the model window and automatic compaction.
Source retrieved 2026-09-28
- ZTABS Yes
Repository-wide migrations are an explicit offering: they decompose legacy PHP, .NET, Java and desktop systems and modernise them module by module over six to eighteen months.
Source retrieved 2026-09-25
The population
143 listings in this index are Published. This is the population as of 2026-10-03. 143 of the 143 published listings in this index carry a settled answer for “Whole-codebase work”. That row asks: Can it carry a migration or refactor across a repository? Every listing in the index is settled on this row, so nothing is left out.
Findings
-
14 of the 143 settled listings answer “Yes” for “Whole-codebase work”. That is 10% of the settled set.
-
38 of the 143 settled listings answer “Qualified” for “Whole-codebase work”. That is 27% of the settled set.
-
11 of the 143 settled listings answer “No” for “Whole-codebase work”. That is 8% of the settled set.
-
76 of the 143 settled listings answer “Does not apply” for “Whole-codebase work”. That is 53% of the settled set.
-
4 of the 143 settled listings answer “Not established” for “Whole-codebase work”. That is 3% of the settled set.
1 of those yes answers is Kilo Code, which still answers yes for “Whole-codebase work”. Its listing carries the stored answer and links to the source recorded for it. This is one worked example, not an independent audit of every source in the census.
What we counted, and how
Each figure above is a count over the “Whole-codebase work” row of the listing datasheet, taken from the same stored answer the listing page renders. The denominator is the 143 listings whose answer is settled, meaning one of yes, qualified, no, does not apply, not established. On this row that is the whole index, because every listing has a recognised stored answer. Every number here is stored with the read-only query that reproduces it and re-run every sixty seconds against the live corpus, so a figure that stops reproducing surfaces as drift rather than as a stale sentence nobody notices. Our full method covers how a datasheet row is settled in the first place.
Limitations
This counts stored datasheet answers, not independently tested capabilities. A sourced answer can record a vendor statement or our reading of published evidence; this census does not re-fetch those sources. “Not established” means we have not established an answer. That can reflect vendor nondisclosure, blocked evidence, or unfinished research, not a no. An absent or unrecognised answer is excluded rather than treated as a researched finding. The figures are restated when the stored corpus changes; the date above is the count used for this published version, not a new verification of the vendors.
Get the next report
New agents rankings and fresh data reports. One short email, one-click unsubscribe.