{"data":{"slug":"zou-group-textgrad","name":"textgrad","tagline":"Automatic 'Differentiation' via Text using Large Language Models to Backpropagate Textual Gradients","github_url":"https://github.com/zou-group/textgrad","owner":"zou-group","repo":"textgrad","owner_avatar_url":"https://avatars.githubusercontent.com/u/30398052?v=4","primary_language":"Python","stars":3700,"forks":294,"topics":["ai-optimization","compound-systems","large-language-models","prompt-optimization","textual-gradients"],"archived":false,"github_pushed_at":"2025-07-25T14:30:39+00:00","maintenance_label":"Dormant","stars_delta_30d":44,"url":"https://www.graphcanon.com/tools/zou-group-textgrad","markdown_url":"https://www.graphcanon.com/tools/zou-group-textgrad.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/zou-group-textgrad","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=zou-group-textgrad","description":"TextGrad: Automatic ''Differentiation'' via Text -- using large language models to backpropagate textual gradients. Published in Nature.","homepage_url":"http://textgrad.com/","license":"MIT","open_issues":66,"watchers":26,"ai_summary":"TextGrad is a tool that leverages large language models to optimize prompts by backpropagating textual gradients.","readme_excerpt":"### Installation\n\nYou can install TextGrad using any of the following methods.\n\n**With `pip`**:\n\n```bash\npip install textgrad\n```\n\n**With `conda`**:\n\n```sh\nconda install -c conda-forge textgrad\n```\n\n> :bulb: The conda-forge package for `textgrad` is maintained [here](https://github.com/conda-forge/textgrad-feedstock).\n\n**Bleeding edge installation with `pip`**:\n\n```sh\npip install git+https://github.com/zou-group/textgrad.git\n```\n\n**Installing textgrad with vllm**:\n\n```sh\npip install textgrad[vllm]\n```\n\nSee [here](https://pip.pypa.io/en/stable/cli/pip_install/) for more details on various methods of pip installation.","github_created_at":"2024-06-11T00:34:56+00:00","created_at":"2026-07-07T17:35:38.407532+00:00","updated_at":"2026-08-18T00:01:42.774427+00:00","categories":[{"slug":"model-training","name":"Model Training","url":"https://www.graphcanon.com/categories/model-training","markdown_url":"https://www.graphcanon.com/categories/model-training.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/model-training"}],"tags":[{"slug":"ai-optimization","name":"ai_optimization"},{"slug":"compound-systems","name":"compound-systems"},{"slug":"large-language-models","name":"large language models"},{"slug":"prompt-optimization","name":"prompt-optimization"},{"slug":"textual-gradients","name":"textual-gradients"}],"trust":{"provenance":{"is_fork":false,"github_id":813368909,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-08-18T00:01:41.576Z","maintenance":{"label":"Dormant","score":18,"methodology":"github_public_v1","releases_90d":0,"days_since_push":388,"last_release_at":"2024-12-15T14:06:27Z","stars_delta_30d":44,"open_issues_delta_30d":0},"security_summary":{"status":"findings","scanner":"osv@v1","low_count":19,"high_count":0,"last_scan_at":"2026-07-11T11:05:51.174Z","medium_count":0,"scan_profile":"deps","critical_count":0}},"capability_facts":{"scan":{"source":"repo_scan","observed_at":"2026-08-18T00:01:42.070Z"},"languages":{"value":["python"],"source":"github.language","observed_at":"2026-08-18T00:01:42.070Z"},"license_spdx":{"value":"MIT","source":"github.license","observed_at":"2026-08-18T00:01:42.070Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["When optimizing complex prompting for large language models in production due to its published effectiveness.","For cutting-edge prompt optimization research, thanks to its integration options beyond just pip."],"when_not_to_use":["If only basic and traditional manual tuning methods are needed for simpler use cases.","Avoid if strict version control is required since the bleeding edge installation points to GitHub directly."],"source":"enrich:decision_facts","observed_at":"2026-07-12T18:53:44.900Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"TextGrad optimizes prompts using large language models to backpropagate textual gradients."}]}}