Xata – the branch native Postgres platform Xata is open source ⭐ Star the repository Xata is open source ⭐ A Postgres branch for every pull request The Postgres platform with instant branching and scale-to-zero compute. Get startedWatch a demo A production Postgres database branches into two active isolated environments while capacity lanes represent hundreds more production-like branches. PR #846 Cart totals Cursor task Add index Preview search-v2 CI migration-228 PR #847 Invoice export Claude task Backfill data Preview dashboard-v3 Agent task Seed catalog PR #846 Cart totals Cursor task Add index Preview search-v2 CI migration-228 PR #847 Invoice export Claude task Backfill data Preview dashboard-v3 Agent task Seed catalog PR #847 Invoice export Claude task Backfill data Preview dashboard-v3 Agent task Seed catalog PR #851 Webhooks Preview account-settings CI schema-check Copilot task Normalize names PR #847 Invoice export Claude task Backfill data Preview dashboard-v3 Agent task Seed catalog PR #851 Webhooks Preview account-settings CI schema-check Copilot task Normalize names PR #851 Webhooks Preview account-settings CI schema-check Copilot task Normalize names PR #854 Audit log Preview pricing-v2 Agent task Optimize query PR #857 Team roles PR #851 Webhooks Preview account-settings CI schema-check Copilot task Normalize names PR #854 Audit log Preview pricing-v2 Agent task Optimize query PR #857 Team roles PR #854 Audit log Preview pricing-v2 Agent task Optimize query PR #857 Team roles PR #846 Cart totals Cursor task Add index Preview search-v2 CI migration-228 PR #854 Audit log Preview pricing-v2 Agent task Optimize query PR #857 Team roles PR #846 Cart totals Cursor task Add index Preview search-v2 CI migration-228 ProductionPrimary Postgres 18.3 Production-like data Healthy99.99% uptime1,324 branches PR #842 Add billing Full copy ready2.2s PR #843 Update schema Full copy ready2.1s PR #846 Cart totals Cursor task Add index Preview search-v2 CI migration-228 PR #847 Invoice export Claude task Backfill data Preview dashboard-v3 Agent task Seed catalog PR #846 Cart totals Cursor task Add index Preview search-v2 CI migration-228 PR #847 Invoice export Claude task Backfill data Preview dashboard-v3 Agent task Seed catalog PR #847 Invoice export Claude task Backfill data Preview dashboard-v3 Agent task Seed catalog PR #851 Webhooks Preview account-settings CI schema-check Copilot task Normalize names PR #847 Invoice export Claude task Backfill data Preview dashboard-v3 Agent task Seed catalog PR #851 Webhooks Preview account-settings CI schema-check Copilot task Normalize names Engineering teams use Xata for agentic development Agentic development metrics Branch creation ~3 sec Growth in branches created 5x MoM Created by coding agents 50% Branch creation cost 1,000 for $1 Modern development needs more than a single staging database. Coding agents multiply the number of changes happening at once. A single shared database cannot safely support all those migrations, tests, and experiments. Shared staging puts four workflows in one database, causing migration conflicts, failed tests, overwritten data, and waiting. Xata creates four isolated Postgres branches with production-like data so every workflow can run independently without conflicts. SHARED STAGING SHARED STAGING Shared staging database Everyone works in the same database Claude — Fix migration Migration conflict PR #842 — Add billing Tests failing Preview #319 — Checkout Test data overwritten PR #845 — Add permissions Waiting for database Schema changed while tests were running 1 database · 4 workflows · Changes collide Isolated Postgres branches Every workflow gets its own branch Claude — Fix migration Branch ready · 2.1s PR #842 — Add billing Tests passed Preview #319 — Checkout Preview ready PR #845 — Add permissions Running tests 4 isolated branches · No conflicts · Created automatically BenefitsChange schemas safely Run migrations and test schema changes without affecting production or anyone else’s work. Test with production-like data Catch issues that empty databases, fixtures, and mocks cannot reveal before changes ship. Work in parallel Give every developer, agent, and preview an isolated database without waiting on shared staging. Clone any Postgres database in seconds Create isolated branches from production without copying the full database. Each branch gets its own schema, compute, and production-like data. Copy-on-write storage Using copy-on-write storage, each branch shares data with its parent and stores only changes instead of creating full copies. Developmentprod-clone Development /rl/episode-001 Development /rl/episode-002 Development /sim/load-test Development /sim/regression Scale to zero Run branches only when needed. When idle, they hibernate and use no compute. 280 branchesActive: 280Idle: 0 Compute$33.60/hr Wake on the first query Branches wake on first query and respond instantly: no connection timeouts, no spin-up delays. > SELECT * FROM users LIMIT 3; Production /feature/auth idnameemail Upstream Postgres Branching happens in the storage layer, not through a custom Postgres fork, so you keep standard PostgreSQL, extensions, drivers, and tooling. address_standardizer address_standardizer_data_us amcheck auto_explain autoinc bloom btree_gin btree_gist citext cube dblink dict_int dict_xsyn earthdistance file_fdw fuzzystrmatch hll address_standardizer address_standardizer_data_us amcheck auto_explain autoinc bloom btree_gin btree_gist citext cube dblink dict_int dict_xsyn earthdistance file_fdw fuzzystrmatch hll address_standardizer address_standardizer_data_us amcheck auto_explain autoinc bloom btree_gin btree_gist citext cube dblink dict_int dict_xsyn earthdistance file_fdw fuzzystrmatch hll address_standardizer address_standardizer_data_us amcheck auto_explain autoinc bloom btree_gin btree_gist citext cube dblink dict_int dict_xsyn earthdistance file_fdw fuzzystrmatch hll hstore hypopg intagg intarray insert_username ip4r isn lo ltree moddatetime pageinspect pg_buffercache pg_cron pg_freespacemap pg_hint_plan pg_partman pg_prewarm hstore hypopg intagg intarray insert_username ip4r isn lo ltree moddatetime pageinspect pg_buffercache pg_cron pg_freespacemap pg_hint_plan pg_partman pg_prewarm hstore hypopg intagg intarray insert_username ip4r isn lo ltree moddatetime pageinspect pg_buffercache pg_cron pg_freespacemap pg_hint_plan pg_partman pg_prewarm hstore hypopg intagg intarray insert_username ip4r isn lo ltree moddatetime pageinspect pg_buffercache pg_cron pg_freespacemap pg_hint_plan pg_partman pg_prewarm pg_repack pg_surgery pg_trgm pg_visibility pg_walinspect pgrowlocks pg_stat_statements pg_stattuple pgaudit pgcrypto pgrouting pglogical pgtap plpgsql postgres_fdw postgis postgis_raster pg_repack pg_surgery pg_trgm pg_visibility pg_walinspect pgrowlocks pg_stat_statements pg_stattuple pgaudit pgcrypto pgrouting pglogical pgtap plpgsql postgres_fdw postgis postgis_raster pg_repack pg_surgery pg_trgm pg_visibility pg_walinspect pgrowlocks pg_stat_statements pg_stattuple pgaudit pgcrypto pgrouting pglogical pgtap plpgsql postgres_fdw postgis postgis_raster pg_repack pg_surgery pg_trgm pg_visibility pg_walinspect pgrowlocks pg_stat_statements pg_stattuple pgaudit pgcrypto pgrouting pglogical pgtap plpgsql postgres_fdw postgis postgis_raster postgis_sfcgal postgis_tiger_geocoder postgis_topology prefix refint seg sslinfo tablefunc tcn tsm_system_rows tsm_system_time unaccent uuid-ossp vector wal2json xml2 postgis_sfcgal postgis_tiger_geocoder postgis_topology prefix refint seg sslinfo tablefunc tcn tsm_system_rows tsm_system_time unaccent uuid-ossp vector wal2json xml2 postgis_sfcgal postgis_tiger_geocoder postgis_topology prefix refint seg sslinfo tablefunc tcn tsm_system_rows tsm_system_time unaccent uuid-ossp vector wal2json xml2 postgis_sfcgal postgis_tiger_geocoder postgis_topology prefix refint seg sslinfo tablefunc tcn tsm_system_rows tsm_system_time unaccent uuid-ossp vector wal2json xml2 Automate every branchCreated when work starts. Removed when it ends. Connect Xata to GitHub, CI, or your coding agents. A Postgres branch is created automatically for every task, pull request, or preview and removed when the work is complete. Automated Pull requests Create a production-like database branch when a pull request opens. Run migrations and tests against it, then delete it after merge. PR #842 Pull request opened Branch created Tests passed PR #842: Pull request opened, branch created, Tests passed. Automated Coding agents Let agents create isolated branches through the API or MCP, safely modify schemas and data, and clean up when the task finishes. Claude task Code session started Branch created Migration applied Claude task: Code session started, branch created, Migration applied. Automated Preview deployments Connect every preview to its own Postgres branch so the complete application, code, schema, and data, can be validated before publishing. Preview #319 Preview deployment Branch created Preview live Preview #319: Preview deployment, branch created, Preview live. How to get startedAdopt Xata without moving production. Keep production where it is, or move it to Xata. Either way, every pull request and deployment preview gets its own isolated Postgres branch. Path 01Keep prod where it is Create a clone of your production database in Xata, keep it in sync with production, and have it ready for instant branching. Branches from a production clone Production data with built-in anonymization Continuously synced with production Create a production clone Path 02Run production on Xata Run your production database on Xata with native branching, built-in high availability, read replicas, and automatic scale-to-zero. Native branches from production High availability Read replicas Create a production database Production-ready Postgres Xata runs vanilla PostgreSQL with high-performance storage, isolation, and instant branching built in. Xata includes built-in production observability for your database — uptime, query latency, throughput, and active connections. xata-dashboard — production All systems operational Last 90 days Uptime99.99% p99 Latency4ms Connections0 Throughput12.4k qps reads writes Vanilla Postgres No forks and no compatibility issues. High availability Automatic failover to read-replicas. PITR backups Restore your database to any point in time. Premium storage Low latency and high throughput storage. ExampleWhat does a database per pull request cost? Consider a company creating 100 pull requests per day against a 100 GB production database. With traditional database infrastructure, giving every pull request its own isolated database quickly becomes unsustainable. Traditional Database Per database compute 24 hrs Per database storage 100 GB Total compute 100 × 24 = 2,400 hours Total storage 100 × 100 GB = 10 TB With Copy-on-Write & Scale-to-Zero Per database compute ~ 5 min Per database storage 3 GB modified Total compute 100 × 5 min = 9 hours Total storage 100 GB + (100 × 3 GB) = 400 GB Storage0.0% less 10 TB 10 TB 400 GB modified 400 GB modified Compute0.0% less 2,400 hrs 2,400 hrs 9 hours 9 hours Built for platformsAdd branch-native Postgres to your product Give every user, workspace, preview, or coding agent an isolated Postgres branch, created through one API and managed by Xata. Provision instantly Create a new Postgres database in a second through a simple API. Branch at any scale Create a production-like branch in seconds, regardless of database size. Scale to zero when idle and wake in about a second. Scale to millions Support large numbers of short-lived databases without duplicating storage or paying for idle compute. Deploy on your terms Use Xata Cloud or run in your own cloud with BYOC, private networking, and regional control. Explore Xata for platforms What our users are saying See what engineers, builders and leaders are sayi…