Measure agent quality with evals

No dark-mode styles — this email renders the same in a dark client.

Run offline evals to ensure agent quality before shipping.

LangSmith (https://dxxRf404.na2.hubspotlinks.com/Ctc/JB+113/dxxRf404/VWBgw03zNQk8W6WgZ0y7j3Vw3W4xGMq_5Wc5z9N1YBnwn3pyd0W95jsWP6lZ3kZW1ZGGSs1R0g_FW4gzL-J3HmB1NV3Gb604HjgjtW2ThxPh4yC_kRW2XwHpj2FYk46W6nb4X98Fq0RmW67f38h6FR5jRW2MPXjv3788HXW3KM-x34wcxTQW6fVW1210ydxMVm-1wv6b2gW7W4QJ-Q072mSCwW49zL_x2x8qbtVScqh75mCbkvW5Fckdf1lntgWW8pgzsM1T0zBWW6ZnnYW2jpp7QN15QL15F2MJsW6Sq8fL3mWC8KW8GJddv1pFGqsN1qmNF5_Y37RW6n_-hH2R-hDGW3jKP0B4w3nmLW4jJQ-v4k_j3cW4WCpLW5wdj77W8SmXh325mNCFW2C85dF8wQhnjW9lWPWX51fT-gN4zb6k6YY52BW92-G4W6ZXsYdd2T6yv04 )

TEST

Measure agent quality before you ship

Test (https://dxxRf404.na2.hubspotlinks.com/Ctc/JB+113/dxxRf404/VWBgw03zNQk8W6WgZ0y7j3Vw3W4xGMq_5Wc5z9N1YBntC5n4LbW5BWr2F6lZ3pbW3MHRyh2FhH3tW2nHnwY7lsY6lN1-6yK8SG7HWW34ynZl61Hjk7N6wdGyfGDSZLW2r3P-m4Qwzp5W6t9GxP3GJjvnVLCdLc8rZT-tW1l0qPs6KpLpdW6Vc2Y05L5CDvMMcZLjJzCkTW8sx5cQ35C4tKW5phJNl6DZ2FLW51yx6X6Y9m2VN3BS8r33bz6-W8YP2x37D18XYW98p-F23kSHkXW2vTmJy5hblYRW46xLYL3zcs-wW8zmq1s40MyhWVcw2Vp2Wb-cBW7b54342jbS1gVWB7Cp3_3p-wW7xHpsl8dPzyFW2kVQC45ShtL0W1vKvm67vCYTZF3_nw9Rh2VQW3gPfQj7Tw5-kW5JFJdD22gnfKW4vnZYK8x3N_ZW4WJN7L1QflxCN83XBLKhJ-CtW7RT5PD6LlhY7W6k2s678vFtLxdDmYds04 )

Know your agent works before you ship it. Build datasets of expected answers, run offline evaluations to check correctness, and track results as experiments.

Create offline evaluator →
(https://dxxRf404.na2.hubspotlinks.com/Ctc/JB+113/dxxRf404/VWBgw03zNQk8W6WgZ0y7j3Vw3W4xGMq_5Wc5z9N1YBntj5n4LbW50kH_H6lZ3pYW65flQb7F-2zPW5GbzqF9jB3gTW14Yhsb99NQs4W97_Vbd4GRLbGW6TxqxR6FJjn4W6zZJSj87nPN3VqVfcm6-nm4GW7WbDD64MFQP8W7Cs2LJ3_1cV-W5j0r4J1FhHT8W7_V0k227jdH2W4lvS-y8GDw3qW7hRdGy7Pm_ynN59gshx72Q-MW1fg7tt8lWxM8W2Vs8hT8WYDx3V_QrQF46D481W7wqtcM6XWq41N6Nz9dYqDbgTW55YGGR1fyBFDW4rY1jp3fw3_LW64XRVY6dZpnLW5zyhkW1-PsLgW51PRFn4JqRv8W8_J8sq8192ydV2_4n_8q9HpZN7PMDkkRK1L5W3KlQ4q6SgByNW6xv1Rt3rrYz8W2LfTG13R_9JGW6zwGGb3rDWXzW56BY07216_v9f3PYnwg04 )

1. Offline evaluators

Run your agent against a curated dataset before you ship, capturing outputs, evaluator scores, and traces so you catch regressions and compare versions ahead of production.

Read docs › (https://dxxRf404.na2.hubspotlinks.com/Ctc/JB+113/dxxRf404/VWBgw03zNQk8W6WgZ0y7j3Vw3W4xGMq_5Wc5z9N1YBntW5n4LbW69t95C6lZ3mhW6cfRtk74SdlhN7v28txDtXSZW2C7bRV2Xbrb0W8zS1g71PcwWqN4nfyKC3D-D1W3mPN7p8ntcZNV24w998MZZbKW3JJDkw42j-tXVy4t9n9fhzz-W3B34vj8C41QKW58-01F2S-l3pW14xfzz91mHfcW62v1387sbh2sW3qh86z81tDfJW8D1Kpp6JlQhMVBY9sk1ZFKvqW8KFYQZ97MQ1mN95bmsHMyDgXW7jTxKr2kRl0MW962JPt9cZy-lW5Mr3wW4q21syW4LD2QL56n4mcW2LLqrM97mmKFW2j5-8q8htg6sW5Gt2xJ2szb2QW5s8rHg1g7Wb8W8KgQkl8ZGb0KW4RlqJb7Nh629W1JRbC664t-KXN15CbbCthp6kW9213tQ5wR_8NW2WrBNT8jdplhVdZ-zN2LVLw6V-DK6k3k3Fw0W5WDJq95F_kCBW8xRZR17SdWzSf95jW0j04 )

2. Datasets

Collections of inputs and (optionally) expected outputs that help evaluate your app's performance.

Read docs › (https://dxxRf404.na2.hubspotlinks.com/Ctc/JB+113/dxxRf404/VWBgw03zNQk8W6WgZ0y7j3Vw3W4xGMq_5Wc5z9N1YBnvb5n4LbW6N1X8z6lZ3nJW4DWsgp5slyWhW54NvzZ4q_XGNW4f8LY627W_khW8D8B5r2QT-KfW6SLmBY48L4dzW8cyg5M1YmPlvW6Ntym36CNWGXW5gtlpF5WrGzRW2vHk3D2rhC6pW79L34G8HQ2mfW5CQkb42YzC0DW8Y_yD11vDdxbW7DcQwz8md05SW487Xcr4-hq1NW680jjK18z5LJV7jXxP7klrg1W6cD_vH6JtCK-M5Zp_cdy2hfW3jd-mf8CM11FW5TmTDj2ZNr52W7VYJVb1Vm3MfW1cVCNx3WxQfnW22R4kN44DLPgW3PxjTb3x7j84W17qVKq3xTrRMVCDzYF1Ycy75W90Kn6L6QScvyW2-xcZ93XTXbxW3FTKXt3jtqs9W3HGPDd4sxtGLW2LDKjX1kGPgkW34VHD42dRfsZW1tMP6m1Mr5Z4W97mLrZ2ZM674N2_KFM2j_5Y7W6ZNTpQ1rbrD2W90Yv_y5zwgZ4W5CmFTH5W9bWhf8s6mcY04 )

3. Experiments

Captures an agent's outputs, evaluator scores, and traces on a dataset, so you can see whether a change actually helped.

Read docs › (https://dxxRf404.na2.hubspotlinks.com/Ctc/JB+113/dxxRf404/VWBgw03zNQk8W6WgZ0y7j3Vw3W4xGMq_5Wc5z9N1YBntW5n4LbW69t95C6lZ3lRW34HMsV7FxfPPW4yxLd84dsF67W65CH5k6V3tKhW2YdCP66-WBtcW6zz4LH5zxhh6W7GZKFm5NV2WlN39FTkdXhT7ZW1Wbrns2WfRRJN513k0YRnMShVxPZqx2GWCp6W7ggt623nFYZPW4Txqkd5mnqZQW8x1Xg95dPGN2W6qHt3k2YSZ-DN5ZTXF2PSpDLW4XkKfv6-y-bFW7fW1Xr3_vtwcW1v-ncb16WLTgW5dfM6d2lkX66N7CPVShkRKg0W48FCwz5NgfFpW734ScT3wHVfPW8DRxC88FJ9ZcW4C0z-V70g_-LW8yZfPs3bktrVN24-4Txb1gc1Vtb4m58f2Sm7W6T2Gvf2YTxGYW2gxHQS3lW7rqW80pvCX7ZF32fW3PWJRk8Shh6rW1tW_Bx5--n6nW29tbvM4N76F5W7Bl3w-3Qf5n2W7xFRYr2K9X09W4FVxLY2n44Z_f3wZKB604 )

Need help?

💬 Support Portal: (https://dxxRf404.na2.hubspotlinks.com/Ctc/JB+113/dxxRf404/VWBgw03zNQk8W6WgZ0y7j3Vw3W4xGMq_5Wc5z9N1YBnwn3pyd0W95jsWP6lZ3pqW3Sb8wm7kggc8W25vXCR1djf7TW5fDqQv4TJDg-W6BztNH19tlwLW61qZ961683cxW20TBY_4BtM2rW4RKZqM6MkLrpN1M-jKmnmndYW6Tljq660D2cJW87-4jJ58fX8hW3cwWKb998368W2WYQRh39DJ15W5p5zF76Ht-mcW16Xcr23qLFyRW74rndp91DmskMD_78xjPXRQN7ZzRFkdpv-bW3FMfsn7SPh1BN1rKz-P6W9mGW8MX7nB3HMgdWW60b6tY3JSZ6RN5rLHG53LSCsW1DrDL54NqTgjW9dfx_j7r6VVnW5gPy0l814XTyW9l1wq62mhZ-6W1hH4V13VL045W71Vldj1c2JwLW5WmHtq6fyjWSVfYz-64cjTTdf91HZ6004 ) Find answers, submit a ticket, or chat with our AI assistant.

📚 Docs (https://dxxRf404.na2.hubspotlinks.com/Ctc/JB+113/dxxRf404/VWBgw03zNQk8W6WgZ0y7j3Vw3W4xGMq_5Wc5z9N1YBntC5n4LbW5BWr2F6lZ3pbW3MHRyh2FhH3tW2nHnwY7lsY6lN1-6yK8SG7HWW34ynZl61Hjk7N6wdGyfGDSZLW2r3P-m4Qwzp5W6t9GxP3GJjvnVLCdLc8rZT-tW1l0qPs6KpLpdW6Vc2Y05L5CDvMMcZLjJzCkTW8sx5cQ35C4tKW5phJNl6DZ2FLW51yx6X6Y9m2VN3BS8r33bz6-W8YP2x37D18XYW98p-F23kSHkXW2vTmJy5hblYRW46xLYL3zcs-wW8zmq1s40MyhWVcw2Vp2Wb-cBW7b54342jbS1gVWB7Cp3_3p-wW7xHpsl8dPzyFW2kVQC45ShtL0W1vKvm67vCYTZF3_nw9Rh2VQW3gPfQj7Tw5-kW5JFJdD22gnfKW4vnZYK8x3N_ZW4WJN7L1QflxCN83XBLKhJ-CtW7RT5PD6LlhY7W6k2s678vFtLxdDmYds04 ) :  (https://dxxRf404.na2.hubspotlinks.com/Ctc/JB+113/dxxRf404/VWBgw03zNQk8W6WgZ0y7j3Vw3W4xGMq_5Wc5z9N1YBnwn3pyd0W95jsWP6lZ3m8VSgljf8hvj9_VNsp9L2D8x6hVx2-JR7-sxxrW8F6W3t6pSj9MW1Kw0HM4LQ_l7W7XsTlV6YMFw4W5_9jRk26_XtqW6g25P16tthwtW4Lkc7r1z-Gz3W1rhJ2v3GPMmLW4QRN6z4CsHR8W6cRjWd9fSvnRN7Wg--xjClGvVMCXgb3t5Qh3W6dZjrt6Zn_Z7W11Lf6D5VHWpcW4yxDv33KKZ5bW1y42xj17NVmrW2ZnSR127dFPWW3gtjh58W6qkGW3Y0Qjp29vFhPN6qBCyd-FMKHW7lQcKY3jMfGqW5BFdYT9fJGYgW3LZJmg8NpnbKW717mXG6mZDh1W48k4k731xHrPW61rSwN6fhdbDW5qXTR01hLMHNN7-3fRrXyWBjf8Hd_Pj04 ) Explore all testing tools within LangSmith.

🎥 Academy (https://dxxRf404.na2.hubspotlinks.com/Ctc/JB+113/dxxRf404/VWBgw03zNQk8W6WgZ0y7j3Vw3W4xGMq_5Wc5z9N1YBntC5n4LbW5BWr2F6lZ3pWW3sGpPC9kRhGKW3JR_KY8-ncZxW5YSBHy3MLSk2W4_rSbW6f70tsW8n2g3y8VwVm4W2zLWF31QCpbxN4W9ly-tWX8tW7jQzBW9jGYHtW8jL_kr6L-XGZW8X5VYh5rDk-CW5lQHNx3XSTRkW7--zm318c9smW7fnkCc4NtnZkN1mx5T88-j_yVCZKPn2KZmTFW5Q7PMW4vH_dwW7g5Dvt1T0LfBW5FNQ0j6Zk5zlW1ds3vW6dNfXJW8h6x0w3S5LRdW98xjdg4Tz1NTW68pxfF1sbdD2W6xHpTj8qwsnGW1-Xp3v2XdppzVKXWZk3Mz2PbW3mKlvR6NGt2BN9dyJlKvblx-W3yCfys5T9BbTW5tmqXP20yNH4W27R2JT9kQ-6lN1ZgZDfr_1bCW76W-zd7r6jh5W62_5c53-j4GYW99rJfb904x9qf6gjRsC04 ) : Enroll in Intro to Agent Observability & Evaluations.

LinkedIn (https://dxxRf404.na2.hubspotlinks.com/Ctc/JB+113/dxxRf404/VWBgw03zNQk8W6WgZ0y7j3Vw3W4xGMq_5Wc5z9N1YBntj5n4LbW50kH_H6lZ3n1W8hVjJz88FbXLN4hF9sQdMygRW6CzRC273zQyCN3K_fBrYQx7NW4WQq2_8lcDHyW6xn9tm317_DpN8mMD1wbCS0nW1s1mzn6m0RjVW96MdGz8SWsl5W9hWq-l8FLqfFW92jPhM6dKKCfN3nf9zLmVQydW6xTF_85T3gTCW17f_h-8RqjrgW4zfXgW1-FZrCW5KZDZ25djznyW9gJxkc4s0MCKW8RrCWk290hfJN13hY7xwWND-W68RDWL6VNc91W7k-NsL3JMQT5W4XxcQQ3vsQGtW80kPKh3p42b_W3PdjwC1bKZfnW6MbxGX2SZKtJW95N8Cy6xwFlNW2zJMRF15tpmZW6tHCvs3NVry_W7KLtFC3r7C02W8NbY982Msm_WW8c9-Ld8LGPXjW2Ltmlk8VL4VJf6Nq7tz04 )

X (https://dxxRf404.na2.hubspotlinks.com/Ctc/JB+113/dxxRf404/VWBgw03zNQk8W6WgZ0y7j3Vw3W4xGMq_5Wc5z9N1YBnwn3pyd0W95jsWP6lZ3mqN89M7PmJgStdN3R2l5YC3vzyW4s0-q758mnlRW4ZVSGp3Zzl0cW3JxnFT82dNM-W61qbzb2cyBt6W1H8Jh216VlPQW51CWXm21r759W2Pzd6N8hPHWHW407NfV6DQX28W47s3XJ9bwV7YW3tpCxR4kb3rnW8QmDCv5DnR8_W4G98H96Llk63W9j2jWN3PLW4tW92_JQf1pKzbyN1KSG46TD7-NW70hL054zBQjgW4X39PL7s-KS0W5x4f_n8_H_XfW7FntvH5bmp-6W2rWYl23Db4MYW3R2Lp35TbNRBW3v3nD_8zYlySW4LvQg92flG42W14myxJ53cm9bW1BFYcd8Zy8qbW4VVm4_7Y9MKJW2NMqty6S0N59W6T8pPw5s9Mwpf10p3v004 )

YouTube (https://dxxRf404.na2.hubspotlinks.com/Ctc/JB+113/dxxRf404/VWBgw03zNQk8W6WgZ0y7j3Vw3W4xGMq_5Wc5z9N1YBntj5n4LbW50kH_H6lZ3p3W4fFZ9j39SMhBW87yN2j4Yrl9nW7KX4JC688Q2YW8GZ3L26lyt0xW58gtWf54JXmWVgCgrf52kxYDW5zkJ5c6Ly7X4W4wD6xj5kND-JW2SDLJ07VzR5nW4T5ckx88KVkqW4XVzYs8JqKj3W76bSZQ5JDlSLW4lqfRZ6XC_ZbN1c4b_Vh4SR5W8g_CbG6dKsZSW9b8rw75jJ1q2N1MWxj9wc3s-VDQXp94Z-ssDW5WLMT42MY1SZW2H9d1Y7VlRTkW8KG08Z2nv5SNW42lnWX1J5GfDN908n-SVjqFkW4N7mKF997MdnW2Skx7R21TkdVW4wsNcz4KpWGTW7hp_qQ8QHXT5VgfV9p1MBt0jW3JLyHr3wdlJ2W6mtn9v3Vh4LHW8yB7GL5PNdCtW4NGfXb1nmzfldFJM3K04 )

LangChain, 2261 Market Street, STE 22961, San Francisco, California 94114, United States

Unsubscribe (https://hs-242623570.s.hubspotemail-na2.net/hs/preferences-center/en/direct?data=W2nWGVW47zkgjW3gn0pB3zbVfJW2Ws3VS3Q-zVlW2r3xVy368hgYW2TzvBL3DZBbzW4cMk7_3gcDRVW3LP5ZQ4r7vYLW2MQ_R04hwvQqW30mJlg1Nh9Y0W2zNQxH4cRvqhW3bs_dN3zhVMHW1Z9brg2zDT7vW4t8z3T2sSLdjW3ZM2FX4fpQy7W3S__qq3bbrHbW3P8q154pxZrRW1ZkVqH1_vtt7W4pDqTs30cmH6W30t6Yq45G51YW2sxHtR2MMkHWW4hhmV841JCPfW2KBmqy3bywtRW3jt96D47LvkHW3DZY922KVCLVW3R5GnW2CXXT1W3XVRMr2KNpxDW1X6vLg1QcF90W1Vrhgl2B2NLwW21jf9r4kmNlgW3VWR7X2nNv5FW348TWc3JX2XjW2MNbLS3SBBGcW38kh_12-nlsZW3B-ZnS2F_Tj8W2t1fCp2CxhQ1W3T4-Mr2CtP3-W2xKzG71NDtcLW32xf6g2WLvSLW1Sh91r34l2StW2PwgwY1QrKX9W2vPZ9-47Ddf1W2Ft7Gf2HTZK8W4rqsfR4tBp5qW251gq83NVmYFW49CDLJ3C6rBqW3_SL2y4mrLb9W30JCRd30dFfDW47Tyfn2WGLTZW2w1Z4p3Fb00wW1XftTW3ddkl0W3z99Dw383x7FW4pqhN622TMzKW2RDf302CMg1HW2y0Wbd36hq92W3jcWd01ZkxsgW49DYBw2CXRT7W45Sv5R45QVQWW49vPjb4chpXWW2MRxg92r4pByW3Vz6bs3SMN3CW2vPCRm2w1VjFW2HHzWn2vyGXSW4pbzg5257XpMW41GBtw3__9Lr0&_hsenc=p2ANqtz-8mfXfNeDeYCvmcr8j3_Z1Q09s2yYwtzQ2zjk1GOavy5YLPnioLsJNPUNMZIVx4LIVEqM89lmswBBOREzFl9_igVFGxXytl6mck-ZwZkREW6VBVH0_iR2Ep6AESKNNmVuIZ6tDT&_hsmi=32112171 )
Manage preferences (https://hs-242623570.s.hubspotemail-na2.net/hs/preferences-center/en/page?data=W2nWGVW47zkgjW3gn0pB3zbVfJW2Ws3VS3Q-zVlW2r3xVy368hgYW2TzvBL3DZBbzW4cMk7_3gcDRVW3LP5ZQ4r7vYLW2MQ_R04hwvQqW30mJlg1Nh9Y0W2zNQxH4cRvqhW3bs_dN3zhVMHW1Z9brg2zDT7vW4t8z3T2sSLdjW3ZM2FX4fpQy7W3S__qq3bbrHbW3P8q154pxZrRW1ZkVqH1_vtt7W4pDqTs30cmH6W30t6Yq45G51YW2sxHtR2MMkHWW4hhmV841JCPfW2KBmqy3bywtRW3jt96D47LvkHW3DZY922KVCLVW3R5GnW2CXXT1W3XVRMr2KNpxDW1X6vLg1QcF90W1Vrhgl2B2NLwW21jf9r4kmNlgW3VWR7X2nNv5FW348TWc3JX2XjW2MNbLS3SBBGcW38kh_12-nlsZW3B-ZnS2F_Tj8W2t1fCp2CxhQ1W3T4-Mr2CtP3-W2xKzG71NDtcLW32xf6g2WLvSLW1Sh91r34l2StW2PwgwY1QrKX9W2vPZ9-47Ddf1W2Ft7Gf2HTZK8W4rqsfR4tBp5qW251gq83NVmYFW49CDLJ3C6rBqW3_SL2y4mrLb9W30JCRd30dFfDW47Tyfn2WGLTZW2w1Z4p3Fb00wW1XftTW3ddkl0W3z99Dw383x7FW4pqhN622TMzKW2RDf302CMg1HW2y0Wbd36hq92W3jcWd01ZkxsgW49DYBw2CXRT7W45Sv5R45QVQWW49vPjb4chpXWW2MRxg92r4pByW3Vz6bs3SMN3CW2vPCRm2w1VjFW2HHzWn2vyGXSW4pbzg5257XpMW41GBtw3__9Lr0&_hsenc=p2ANqtz-8mfXfNeDeYCvmcr8j3_Z1Q09s2yYwtzQ2zjk1GOavy5YLPnioLsJNPUNMZIVx4LIVEqM89lmswBBOREzFl9_igVFGxXytl6mck-ZwZkREW6VBVH0_iR2Ep6AESKNNmVuIZ6tDT&_hsmi=32112171 )

Accession

nse/langchain/2026-10-16/4287

From

LangChain · hello@mail.langchain.com

Sent

16 October 2026 · 18:00 UTC